the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction...

53
1 Comparing SWAT with SWAT-MODFLOW hydrological simulations when assessing 1 the impacts of groundwater abstractions for irrigation and drinking water 2 Wei Liu 1 , Seonggyu Park 2, 3 , Ryan T. Bailey 2 , Eugenio Molina-Navarro 1, 4 , Hans Estrup Andersen 1 , 3 Hans Thodsen 1 , Anders Nielsen 1 , Erik Jeppesen 1 , Jacob Skødt Jensen 5 , Jacob Birk Jensen 6, 7 and 4 Dennis Trolle 1 . 5 1 Department of Bioscience, Aarhus University, Silkeborg, Denmark; 6 2 Department of Civil and Environmental Engineering, Colorado State University, Fort Collins, 7 Colorado, USA; 8 3 Blackland Research & Extension Center, Texas A&M AgriLife, Temple, United States; 9 4 Deparment of Geology, Geography and Environment, University of Alcalá. Alcalá de Henares, 10 Madrid, Spain. 11 5 NIRAS, Aarhus, Denmark; 12 6 Department of Civil Engineering, Aalborg University, Aalborg, Denmark; 13 7 WatsonC, Aalborg, Denmark. 14 Correspondence: Wei Liu ([email protected], [email protected]) 15 Key Points16 We compared the performance of SWAT and SWAT-MODFLOW and assessed the simulated 17 streamflow signals in response to a range of groundwater abstraction scenarios targeted for 18 irrigation and drinking water. 19 The SWAT-MODFLOW complex was further developed to enable the application of the Drain 20 Package and an auto-irrigation routine. 21 A PEST-based approach was developed to calibrate the coupled SWAT-MODFLOW. 22 The SWAT-MODFLOW model produced more realistic results on groundwater abstraction 23 effects on streamflow. 24 25

Upload: others

Post on 08-May-2020

9 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

1  

Comparing SWAT with SWAT-MODFLOW hydrological simulations when assessing 1 

the impacts of groundwater abstractions for irrigation and drinking water 2 

Wei Liu1, Seonggyu Park2, 3, Ryan T. Bailey2 , Eugenio Molina-Navarro1, 4, Hans Estrup Andersen1, 3 

Hans Thodsen1, Anders Nielsen1, Erik Jeppesen1, Jacob Skødt Jensen5, Jacob Birk Jensen6, 7 and 4 

Dennis Trolle1. 5 

1Department of Bioscience, Aarhus University, Silkeborg, Denmark; 6 

2Department of Civil and Environmental Engineering, Colorado State University, Fort Collins, 7 

Colorado, USA; 8 

3Blackland Research & Extension Center, Texas A&M AgriLife, Temple, United States; 9 

4Deparment of Geology, Geography and Environment, University of Alcalá. Alcalá de Henares, 10 

Madrid, Spain. 11 

5NIRAS, Aarhus, Denmark; 12 

6Department of Civil Engineering, Aalborg University, Aalborg, Denmark; 13 

7WatsonC, Aalborg, Denmark. 14 

Correspondence: Wei Liu ([email protected], [email protected]) 15 

Key Points: 16 

We compared the performance of SWAT and SWAT-MODFLOW and assessed the simulated 17 

streamflow signals in response to a range of groundwater abstraction scenarios targeted for 18 

irrigation and drinking water. 19 

The SWAT-MODFLOW complex was further developed to enable the application of the Drain 20 

Package and an auto-irrigation routine. 21 

A PEST-based approach was developed to calibrate the coupled SWAT-MODFLOW. 22 

The SWAT-MODFLOW model produced more realistic results on groundwater abstraction 23 

effects on streamflow. 24 

 25 

Page 2: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

2  

Abstract 26 

Being able to account for temporal patterns of streamflow, the distribution of groundwater resources, 27 

as well as the interactions between surface water and groundwater is imperative for informed water 28 

resources management. We hypothesize that, when assessing the impacts of water abstractions on 29 

streamflow patterns, the benefits of applying a coupled catchment model relative to a lumped semi-30 

distributed catchment model outweigh the costs of additional data requirement and computational 31 

resources. We applied the widely used semi-distributed SWAT model and the recently developed 32 

SWAT-MODFLOW model, which allows full distribution of the groundwater domain, to a Danish, 33 

lowland, groundwater-dominated catchment, the Uggerby River Catchment. We compared the 34 

performance of the two models based on the observed streamflow and assessed the simulated 35 

streamflow signals of each model when running four groundwater abstraction scenarios with real wells 36 

and abstraction rates. The SWAT-MODFLOW model complex was further developed to enable the 37 

application of the Drain Package of MODFLOW and to allow auto-irrigation on agricultural fields and 38 

pastures. Both models were calibrated and validated, and an approach based on PEST was developed 39 

and utilized to enable simultaneous calibration of SWAT and MODFLOW parameters. Both models 40 

demonstrated generally good performance for the temporal pattern of streamflow, albeit SWAT-41 

MODFLOW performed somewhat better. In addition, SWAT-MODFLOW generates spatially explicit 42 

groundwater-related outputs, such as spatial-temporal patterns of water table elevation. In the 43 

abstraction scenarios analysis, both models indicated that abstraction for drinking water caused some 44 

degree of streamflow depletion, while abstraction for auto-irrigation led to a slight total flow increase 45 

(but a decrease of soil or aquifer water storages, which may influence the hydrology outside the 46 

catchment). In general, the simulated signals of SWAT-MODFLOW appeared more plausible than 47 

those of SWAT, and the SWAT-MODFLOW decrease in streamflow was much closer to the actual 48 

volume abstracted. The impact of drinking water abstraction on streamflow depletion simulated by 49 

SWAT was unrealistically low, and the streamflow increase caused by irrigation abstraction was 50 

exaggerated compared with SWAT-MODFLOW. We conclude that the further developed SWAT-51 

MODFLOW model calibrated by PEST had a better hydrological simulation performance, wider 52 

possibilities for groundwater analysis, and much more realistic signals relative to the semi-distributed 53 

SWAT model when assessing the impacts of groundwater abstractions for either irrigation or drinking 54 

water on streamflow; hence, it has the potential to be a useful tool in the management of water 55 

resources in groundwater-affected catchments. However, this comes at the expense of higher 56 

computational demand and more time consumption. 57 

Page 3: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

3  

1. Introduction 58 

The interaction between groundwater and surface water is an important aspect of the water cycle, and 59 

the management or use of one often impacts the availability and temporal patterns of the other. 60 

Improper management and over-exploitation of these water resource components influence the 61 

sustainability of both the water resource itself and also the ecosystems that it supports. Groundwater 62 

abstraction can cause a decline of the water table, and it thereby directly affects surface water bodies 63 

connected to the aquifer (Jeppesen et al., 2015; Vainu and Terasmaa, 2016; Stefania et al., 2018). For 64 

rivers in which a considerable portion of the streamflow is base flow, this can have a strong influence 65 

on the general flow and deteriorate the function of river ecosystems (Johansen et al., 2011; Pardo and 66 

Garcia, 2016). However, interactions between groundwater and surface water are difficult to observe 67 

and measure, and it is, therefore, difficult to determine how much of the reduced streamflow recorded 68 

in some rivers is due to abstractions and how much is due to natural weather-induced variability in 69 

water table elevation. 70 

For quantitative assessment of the impacts of pumping wells on streamflow, a hierarchy of modeling 71 

tools has been developed, ranging from analytical models based on simple water balance equations to 72 

regional, three-dimensional numerical models, depending on the complexity and available data source 73 

of the site (Chen and Yin, 2001; Parkin et al., 2007; May and Mazlan, 2014). Analytical models generally 74 

require less data for parameter identification and may therefore be applied when available data are 75 

sparse, thus offering water managers a simple approach for estimating streamflow depletion with less 76 

time, expertise, and financial costs (Glover and Balmer, 1954; Hunt, 1999; Huang et al., 2018; Zipper 77 

et al., 2018). Nevertheless, as they do not simulate many of the physical processes and ignore the real-78 

world complexity, they may render unrealistic results. In contrast, numerical, process-based models 79 

consider the entire complexity and heterogeneity of river-aquifer systems. Such models can simulate 80 

the regional groundwater dynamics as well as the interactions between groundwater and surface water. 81 

They are therefore part of local water management applications including estimation of streamflow 82 

depletion, although they are generally more time-consuming and costly to set up, calibrate, test, and 83 

apply. 84 

MODFLOW is a  physically-based, fully-distributed, and three-dimensional (3D) finite-difference 85 

groundwater model, and it is considered a state-of-the-art international standard for simulating and 86 

predicting groundwater conditions (http://water.usgs.gov/ogw/modflow/). It can be used to simulate 87 

both steady state and transient conditions. MODFLOW outputs include groundwater hydraulic head 88 

or drawdown at the center of each grid cell as well as groundwater flow rates to/from each stream 89 

Page 4: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

4  

segment if the River (RIV) Package or the Streamflow Routing Package (SFR) is used (Wei et al., 90 

2018). A number of studies have applied MODFLOW to assess the impact of groundwater abstraction 91 

on surface water resources (Sanz et al., 2011; May and Mazlan, 2014; Shafeeque et al., 2016; Stefania 92 

et al., 2018). However, MODFLOW does not simulate surface processes such as land-atmosphere 93 

interactions, agricultural management practices, and surface runoff (Lachaal et al., 2012; Surinaidu et 94 

al., 2014). To obtain spatial-temporal varying recharge rates, MODFLOW is therefore often linked 95 

with land-surface models such as the Precipitation-Runoff Modelling system (Markstrom et al., 2008; 96 

Markstrom et al., 2015) and the Soil and Water Assessment Tool (SWAT) (Izady et al., 2015; Wei et 97 

al., 2018). 98 

The SWAT model is a semi-distributed catchment-scale model and has been widely used to simulate 99 

surface runoff, sediment erosion, pesticide and microorganism transport, and nutrient cycling in 100 

catchments at different geographical locations and scales (Nielsen et al., 2013; Fukunaga et al., 2015; 101 

Malago et al., 2017; Liu et al., 2019). In SWAT, the basin is divided into subbasins through a 102 

topography-based delineation, each subbasin containing a tributary of the river. Each subbasin is 103 

further divided into Hydrologic Response Units (HRUs), which are unique combinations of land use, 104 

soil type, and surface slope. When simulating hydrological dynamics, the areas of the HRUs are 105 

lumped within each subbasin, which makes SWAT computationally very efficient, but this comes at 106 

the expense of losing the spatial discretization of HRUs within a subbasin. SWAT has been utilized to 107 

simulate and quantify the groundwater resources (Ali et al., 2012; Cheema et al., 2014; Shafeeque et 108 

al., 2016) or the effects of drinking water or irrigation pumping on streamflow (Güngör and Göncü, 109 

2013; Lee et al., 2006). However, the SWAT model has traditionally emphasized surface processes as 110 

the model only includes a relatively simple representation of groundwater dynamics, and its output 111 

does not give any spatially explicit information on the groundwater table. In the most recent version 112 

of SWAT (v. 670), groundwater is represented by a lumped module in individual subbasins divided 113 

into a shallow and a deep aquifer. Both the shallow and the deep aquifer may contribute to streamflow 114 

as baseflow through a linear reservoir approximation, ignoring distributed parameters such as 115 

hydraulic conductivity and storage coefficients (Kim et al., 2008). With this simplified implementation 116 

of groundwater dynamics in SWAT, the model can mislead evaluation of groundwater resources or 117 

perform rather poorly in catchments where the streamflow is strongly dependent on groundwater 118 

discharge (Gassman et al., 2014). 119 

To the best of our knowledge, there are two main approaches for making SWAT perform better in 120 

groundwater-dominated catchments. One approach is to modify the SWAT groundwater module code 121 

au566015
Highlight
Page 5: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

5  

itself. For example, (Zhang et al., 2016) modified the subroutines in the SWAT source code by 122 

converting the shallow aquifer water storage change into water table fluctuation with three 123 

groundwater parameters added, namely specific yield, the bottom bed burial depth, and shallow aquifer 124 

porosity. The modified SWAT could then simulate both water table fluctuations and water storage of 125 

the shallow aquifer in time and space. However, it still applied a lumped, linear reservoir approach to 126 

simulate groundwater storage and derive the water table at HRU level, which could give rise to errors 127 

as the HRUs are not spatially explicit within a subbasin. (Pfannerstill et al., 2014) implemented a three-128 

storage concept in the groundwater module by splitting the shallow aquifer into a fast and a slow 129 

contributing aquifer. (Nguyen and Dietrich, 2018) replaced the deep aquifer in the original SWAT 130 

model with the multicell aquifer model. In both of these studies, the modified SWAT model achieved 131 

a better prediction of baseflow than the original SWAT model. However, both models only improved 132 

a part of aquifer system simulation, either the shallow aquifer or the deep aquifer. In addition, they 133 

maintained the semi-distributed approach. 134 

The other approach for improving the performance of SWAT in groundwater-dominated catchments 135 

is to couple SWAT with a physically based, spatially distributed numerical groundwater model, such 136 

as MODFLOW. There are a few studies that have tried to integrate SWAT and MODFLOW code into 137 

one model complex (Kim et al., 2008; Yi and Sophocleous, 2011; Guzman et al., 2015; Bailey et al., 138 

2016). The most recent of these, the SWAT-MODFLOW code developed by  (Bailey et al., 2016) 139 

couples the most recent SWAT code with the MODFLOW-NWT code (a Newton-Raphson 140 

formulation for MODFLOW-2005 (Niswonger et al., 2011), which improves the solution of 141 

unconfined groundwater-flow problems). This coupled version has several advantages over others: an 142 

efficient HRU-grid cell mapping scheme (including generation of geographically explicit HRUs), the 143 

ability to use SWAT and MODFLOW models of different spatial coverage, the use of the 144 

MODFLOW-NWT code, public availability, and a graphical user interface that has been recently 145 

developed for its application (Bailey et al., 2017; Park et al., 2018). Recently, the current published 146 

SWAT-MODFLOW code (Version 2 on the SWAT website) has been applied to catchments in the 147 

USA (Bailey et al., 2016; Abbas et al., 2018; Gao et al., 2019), Canada (Chunn et al., 2019), Denmark 148 

(Molina-Navarro et al., 2019), Iran (Semiromi and Koch, 2019), and Japan (Sith et al., 2019). It has 149 

also been further developed for application in large-scale mixed agro-urban river basins (Aliyari et al., 150 

2019). Within the coupled SWAT-MODFLOW framework, SWAT simulates surface hydrological 151 

processes, whereas MODFLOW-NWT simulates groundwater flow processes and all associated 152 

sources and sinks on a daily time step. In addition, the HRU-calculated deep percolation from SWAT 153 

is passed to the grid cells of MODFLOW as recharge, and MODFLOW-calculated groundwater-154 

Page 6: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

6  

surface water interaction fluxes are passed to the stream channels of SWAT. Hence, the model complex 155 

accounts for two-way interactions between groundwater and surface waters and thereby enables a 156 

potentially much better representation and thus understanding of the spatial-temporal patterns of 157 

groundwater-surface water interactions, which are of key importance to catchment management in 158 

groundwater-dominated catchments. 159 

In Denmark, approximately 800 million m3 of water are abstracted annually and used for irrigation or 160 

drinking water (GEUS, 2009), making the country highly dependent on groundwater. Since the very 161 

dry summers in 1975 and 1976 led to dry out of many watercourses around some cities in Denmark, 162 

the national government has endeavored to regulate the abstraction of surface and groundwater to a 163 

level preventing negative impacts on in-stream biota. Gradually, direct abstraction from surface waters 164 

has been prohibited and groundwater abstraction is regulated to secure a certain minimum flow in all 165 

Danish rivers, mainly by moving the abstraction wells away from riverbanks and wetlands and 166 

implementing a groundwater abstraction permit authority system. However, there still remains some 167 

areas where groundwater exploitation is above the sustainable yield and causes streamflow depletion 168 

according to the national water resource model (Henriksen et al., 2008). 169 

To better understand how abstraction wells used for drinking water or irrigation may influence nearby 170 

streamflow, we applied both SWAT and SWAT-MODFLOW to a lowland catchment in Northern 171 

Denmark – the Uggerby River Catchment. We hypothesize that, when assessing impacts of water 172 

abstractions on streamflow patterns, the benefits of applying SWAT-MODFLOW relative to SWAT 173 

outweigh the costs of additional data requirements and computational resources. We compared the 174 

performance of the two models and assessed the simulated signals of streamflow in a range of 175 

groundwater abstraction scenarios with real wells and abstraction rates for either drinking water or 176 

irrigation with both models. The SWAT-MODFLOW complex used in this study was further 177 

developed based on the publically available version (https://swat.tamu.edu/software/swat-modflow/) 178 

to enable application of the Drain Package of MODFLOW and to allow auto-irrigation. In addition, an 179 

approach based on PEST (Doherty, 2018) was developed to calibrate the coupled SWAT-MODFLOW 180 

by adjusting SWAT and MODFLOW parameters simultaneously against the observations of both 181 

streamflow and groundwater table. 182 

183 

184 

Page 7: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

7  

2. Materials and methods 185 

2.1 Study area 186 

The Uggerby River Catchment lies between latitude 57°17′10"- 57°35′25" N and longitude 9°58′47"- 187 

10°19′55" E. It covers an area of 357 km2 and is located in the Municipality of Hjørring, which is 188 

situated in the northern part of Jutland, Denmark (Fig. 1). The Uggerby River originates from the 189 

southern part of Hjørring in Sterup and winds its way through the area of Hjørring, Sindal, Mosbjerg, 190 

Bindslev, and Uggerby and then discharges into the bay Tannisbugten at the coast of the North Sea. 191 

The study area has a typical Atlantic climate, which is temperate with an average annual temperature 192 

around 8 °C, being warmest in August (17 °C average) and coldest in January (0.5 °C average). The 193 

average annual precipitation during the study period 2002-2015 was approximately 933 mm with no 194 

obvious distinctions among seasons. 195 

Figure 1. 196 

The mean catchment elevation is 34.5 m a.s.l and ranges from 0 to 108 m. Land cover in the catchment 197 

is dominated by agricultural land, and the other land use types include evergreen forest, pasture, 198 

wetland, and urban areas. The soil types are loamy sand, sandy loam, and sand. The main crops grown 199 

in the area include winter wheat, winter rape, barley, corn, and grass. Artificial tile drains have been 200 

installed in parts of the agricultural land in the catchment, although the precise drainage locations are 201 

somewhat uncertain (Olesen, 2009). According to an investigation carried out by Hjørring 202 

Municipality in 2009, there are 101 drinking water pumping wells registered within the catchment and 203 

57 irrigation pumping wells placed on pasture and agricultural land. Generally, irrigation only occurs 204 

from April to October. The average annual irrigation amount varies from 80 to 200 mm depending on 205 

the types of crop and soil conditions (Aslyng, 1983). 206 

2.2 Model set-up and coupling 207 

2.2.1 SWAT model set-up 208 

We used the QSWAT 1.5 interface (George, 2017), which works with the latest SWAT Editor version 209 

2012.12.19 and is integrated into a QGIS 2.8.1 interface. The input data for the SWAT model in this 210 

study include topography, land use, soil, climate, agricultural management, wells, and wastewater 211 

discharge as point sources. 212 

Page 8: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

8  

The catchment was divided into 19 subbasins (Fig. 1) based on the 32 m pixel size Digital Elevation 213 

Model (DEM), which has been resampled from a 1.6 m LIDAR DEM (Knudsen and Olsen, 2008). For 214 

the creation of HRUs, we used the land use map based on the Danish Area Information System (NERI, 215 

2000) and the soil map based on a national three-layer soil map with a 250 m grid resolution (Greve 216 

et al., 2007), and surface slope type was classified into three classes (<2%, 2-6%, >6%). To reduce the 217 

number of HRUs and facilitate the posterior model linkage process, land use for range-grasses and 218 

range-brush, which covered only 1.3% and 1.9% of the total catchment area, respectively, were merged 219 

into pasture, and water (0.9%) was merged into wetland areas. In order to represent the agricultural 220 

management practices in detail, the agricultural area was split into three equally sized types with 221 

different five-year crop rotation schedules (Table 1) based on the real contour of agricultural field plots 222 

and the land use map. Similar to land use, soil types covering a minor part of the catchment (1% or 223 

less) were merged into similar soil types. The distribution and proportion of each land use, soil type, 224 

and slope band after reclassification are shown in Fig. 2. Based on the combination of land use, soils, 225 

and slope, the catchment was discretized into 2620 HRUs. 226 

Figure 2. 227 

Climate data used in the model comprised the 10 km-grid national daily precipitation data (six stations 228 

inside the catchment), 20 km-grid daily solar radiation and wind speed data (five stations inside or near 229 

the catchment), gauged-level daily maximum and minimum temperatures, and relative humidity data 230 

(one station, 27 km from the catchment ) during 1997-2015 from the Danish Meteorological Institute 231 

(Lu et al., 2016). 232 

Farm type and manure/mineral fertilizer application of each agricultural rotation as well as dates of 233 

sowing, harvesting, and tillage were assigned based on reported statistics for 2005 available from 234 

http://www.dst.dk/en (Table 1). We do not know the specific tile drain distribution within the entire 235 

catchment. In general, loamy soils in relatively flat areas are known to be tile drained in Denmark 236 

(Olesen, 2009). To represent this situation, tile drains were set up in agricultural land with a slope less 237 

than 2% and for soil types with a clay content above 8% (Thodsen et al., 2015), representing 27% of 238 

the agricultural land in the catchment. 239 

We assumed that irrigation only occurs in the HRUs where irrigation pumping wells exist (based on a 240 

MODFLOW model created by NIRAS A/S). It is difficult to know the exact dates and water amount 241 

used for irrigation. Thus, to simulate the irrigation, auto-irrigation management was set up based on 242 

heat unit scheduling for the HRUs containing irrigation pumping wells. For the auto-irrigation of crops, 243 

Page 9: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

9  

the water resource used for pumping was defined as the shallow aquifer, and the soil water content, 244 

commonly used as an indicator in actual field irrigation (Chen et al., 2017), was selected as the water 245 

stress identifier with 70 mm as the initial water stress threshold. With the number and location of 246 

pumping wells as well as their pumping rates obtained from the Well Package in the MODFLOW 247 

model, the water abstraction amounts from drinking water wells were added up in each subbasin and 248 

set as the water use pumped from the shallow aquifer in SWAT. 249 

The only significant point source of the study area is the discharge from the wastewater treatment plant 250 

in Sindal located in subbasin 16. With a few other minor sources aggregated to a total discharge from 251 

the wastewater treatment plant, a total of 2768.8 m3 of water was discharged into the stream per day 252 

(data is based on an average for the period 2007-2010). 253 

2.2.2 MODFLOW-NWT model set-up 254 

A steady-state version of the MODFLOW-NWT model has previously been set up for the entire 255 

Hjørring Municipality, covering an area of 930 km2, in which the Uggerby River Catchment is situated 256 

(Fig. 3). The model set-up was firstly established in 2011 and then updated in 2016 by the consultant 257 

company NIRAS A/S and Hjørring Water Supply Company, and has been applied for water resources 258 

management in the Hjørring Municipality. In the model set-up, the geology is represented by 5 hydro-259 

stratigraphic layers, discretized into 183,112 grids (376 rows and 487 columns) with a discretization 260 

of 100 x 100 meters. The uppermost layer is unconfined and the remaining four layers are confined. 261 

The Upstream Weighting (UPW) Package for MODFLOW, which contains hydraulic properties of 262 

each cell, was used as the internal flow package, and a number of boundary condition packages, 263 

including Time-variant specified-head Package, Drain Package, River Package, Well Package, and 264 

Recharge Package, were employed in the model to simulate external stresses. The steady-state model 265 

was calibrated using 1,063 head observations sampled during the period 1996-2010 at 1,006 well 266 

locations distributed within the first, third, and fifth layer by a combination of manual calibration and 267 

auto-calibration through PEST (http://www.pesthomepage.org/). 268 

Figure 3. 269 

Eighteen different hydraulic conductivity values exist in the originally calibrated MODFLOW model. 270 

In order to facilitate the posterior SWAT-MODFLOW calibration, we reclassified and grouped the 271 

specific hydraulic conductivities into five groups. The grouping was made for grid cells of similar 272 

specific hydraulic conductivities, representing the sedimentary materials of clay, silt, silty sand, 273 

Page 10: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

10  

mixture of silty sand and clean sand, and clean sand, respectively. Each group was assigned a unique 274 

specific hydraulic conductivity, which could be targeted for calibration. 275 

For the SWAT-MODFLOW set-up, we converted the modified calibrated steady-state model into a 276 

transient model by assigning values to the specific yield (only for the unconfined layer) and specific 277 

storage of each cell according to the type of sedimentary materials of the cell and representative values 278 

of storage coefficients. The simulated heads generated by the steady-state model were used as the 279 

initial head conditions for the transient model. 280 

2.2.3 SWAT-MODFLOW coupling 281 

SWAT and MODFLOW were combined using the coupling framework developed by (Bailey et al., 282 

2016) and following the procedures described in the instructions available from the SWAT website 283 

(http://swat.tamu.edu/software/swat-modflow/). 284 

For this study, the following changes were made to the original SWAT-MODFLOW code: (1) the grid 285 

cells in the Drain Package were linked with SWAT subbasins so that groundwater removed from 286 

subsurface drains is routed to stream channels; and (2) groundwater pumping in agricultural areas or 287 

pastures is dictated by irrigation applied to HRUs through SWAT’s auto-irrigation routines. For the 288 

latter, this is achieved by calculating the daily volume of applied irrigation water (irrigation depth * 289 

HRU area) and then extracting this volume from the underlying grid cells using MODFLOW’s Well 290 

Package (Fig. 4). When applying the Drain Package of MODFLOW, the original tile drain routine in 291 

SWAT was disabled. The steps in the coupling procedure included: 1) disaggregation of HRUs to 292 

disaggregated hydrologic response units (DHRUs) through GIS processing to make the model spatially 293 

explicit; and 2) creation of six linking text files (HRUs to DHRUs, DHRUs to MODFLOW grids, 294 

MODFLOW grids to DHRUs, MODFLOW river cells to SWAT subbasin rivers, MODFLOW drain 295 

cells to subbasin rivers, irrigation pumping wells in HRUs to MODFLOW grids) through GIS 296 

processing. All related files (MODFLOW input files, original SWAT model files, linkage files) were 297 

stored in one working directory for SWAT-MODFLOW execution. 298 

Figure 4. 299 

2.3 Model calibration 300 

2.3.1 SWAT calibration 301 

Page 11: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

11  

The Sequential Uncertainty Fitting Algorithm (SUFI2), which is implemented in the SWAT-CUP 302 

software (Abbaspour, 2015), was used to calibrate discharge performance in SWAT. The latest 303 

SWAT-CUP version (5.1.6.2) was used. Calibration was performed based on daily discharge records 304 

from 1 Jan. 2002 to 31 Dec. 2008, with a previous 5-year model warm-up period and using Nash-305 

Sutcliffe efficiency (NSE) as the objective function. Five parameters at basin-wide level and 17 306 

parameters at subbasin level related to streamflow were selected and assigned initial calibration value 307 

ranges based on expert judgement and previous SWAT applications in Danish catchments (Lu et al., 308 

2015; Molina-Navarro et al., 2017). 309 

In the study area, two hydrologically connected monitoring stations are found, located at the outlet of 310 

subbasin 13 (station A) and subbasin 18 (station B), respectively (Fig. 1). The two stations represent a 311 

small (average discharge 1.95 m3 s-1) and relatively large (average discharge 4.56 m3 s-1) stream in 312 

Denmark, and both were used for calibration and validation in this study. Station A is located upstream 313 

from station B and its flow therefore has an influence on station B. Thus, the simulated discharge of 314 

station A was preliminarily calibrated first (initial range of related parameters are shown in Table 2), 315 

running 3 iterations with 500 simulations each. After the final iteration for station A, the subbasin level 316 

parameters for the area upstream station A were fixed, while the final ranges of the basin-wide 317 

parameters were used in the subsequent calibration of station B. As the basin-wide parameter values 318 

can impact the hydrology of the entire catchment, for the calibration of station B, discharge data from 319 

both station A and B were included in the objective function. An additional three iterations with 500 320 

simulations were run, where the subbasin level parameters for the remaining area upstream station B 321 

were calibrated using the same initial parameter range as for station A (Table 2), while the basin-wide 322 

parameter ranges from the final calibration step for station A were used as initial ranges. By this 323 

approach, we attempted to make the basin-level parameters representative for both upstream and 324 

downstream areas. Afterwards, the water stress threshold was calibrated manually to ensure proper 325 

simulation of the annual irrigation amount, which ranges from 80 to 120 mm yr-1 and occurs in the 326 

period April to October (Aslyng, 1983). Once the calibration was completed and the parameters were 327 

fixed, we validated the model by running one simulation from 1 Jan. 1997 to 31 Dec. 2015 using the 328 

first 12-years as a warm-up period. 329 

To analyze parameter sensitivity and make the sensitivity analysis comparable with SWAT-330 

MODFLOW, an additional iteration with 500 simulations was run for the calibration period. In this 331 

iteration, the ranges of basin level parameters and subbasin level parameters for the area upstream 332 

station A were the same as those in the final calibration step for station A, while the ranges of subbasin 333 

Page 12: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

12  

level parameters for the area upstream station B were identical with the final calibration step for station 334 

B. 335 

Table 2. 336 

2.3.2 SWAT-MODFLOW calibration 337 

After model coupling, the SWAT-MODFLOW was calibrated by adjusting SWAT and MODFLOW 338 

parameters simultaneously against the observations of both streamflow and groundwater table through 339 

a combination of manual calibration and auto-calibration by the widely used PEST approach (Doherty, 340 

2018). The periods used for model warm-up, calibration, and validation were identical to those used 341 

for SWAT. SWAT-MODFLOW can also be run through SWAT-CUP, whereby the summary statistics 342 

of model performance can be derived and directly compared between SWAT and SWAT-MODFLOW. 343 

In addition, model.in and Swat_Edit.exe, which are included in the creation of the SWAT-CUP project 344 

folder, can be used to adjust SWAT parameters within the PEST routine. 345 

The framework using PEST to calibrate SWAT-MODFLOW was firstly introduced by (Park, 2018). 346 

We applied this framework to this study as well but with BEOPEST (Doherty, 2018) instead of PEST 347 

as the PEST executable file. Figure 5 presents a schematic diagram illustrating how PEST is utilized 348 

for the SWAT-MODFLOW calibration in this study. Five types of files are required to run PEST: 349 

PEST control file, PEST executable file, model batch file, model input template files, and model output 350 

instruction files. The PEST control file is a master file that contains control variables, initial values 351 

and ranges of model parameters, observations and their weights for deriving the value of the objective 352 

function, as well as names of all input and output files related to calibration. BEOPEST was used as 353 

the PEST executable file that enables parallelization of model runs on multiple computer cores, thereby 354 

shortening the calibration time considerably. After each iteration of a PEST run, the PEST optimization 355 

algorithm adjusts the model parameter values to optimize the value of the objective function. The 356 

newly updated model parameter values are then written to model input files using input template files 357 

and Swat_Edit.exe. Next, the SWAT-MODFLOW executable is called by a batch file and generates a 358 

set of output files if the model runs successfully. A python script (exsimvalue.py) extracts the 359 

simulated values from the streamflow output file (output.rch) and the groundwater table output file 360 

(swatmf_out_MF_obs). The extracted simulated data are read by PEST using information from the 361 

model output instruction file and then compared against the corresponding observations. Each iteration 362 

includes a number of model runs according to the control variable set in the PEST control file to allow 363 

adjustment of parameter values. After each iteration, the objective function and a Jacobian matrix are 364 

Page 13: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

13  

calculated, based on which the PEST will make its decision for the next iteration until one of its 365 

stopping criteria, specified in the PEST control file, is met. More detailed information about the 366 

optimization process and principles of PEST can be found in (Zhulu, 2010) and the PEST manual 367 

(Doherty, 2018). 368 

Figure 5. 369 

As shown in Table 3, 26 parameters from SWAT related to surface hydrological processes and 13 370 

parameters from MODFLOW were selected and calibrated through PEST. For SWAT parameters, 371 

with the parameters related to tile drains and groundwater excluded, the final calibrated parameter 372 

values used in SWAT were applied as the initial values in PEST, and the parameter ranges used in the 373 

iteration for SWAT parameter sensitivity analysis were employed as the parameter ranges in PEST. 374 

By manually adjusting MODFLOW parameter values to test their impacts on model outputs, storage 375 

coefficients (SY and SS), horizontal hydraulic conductivity (HK), and two drain conductance (COND) 376 

were deemed as the potential sensitive parameters, with the value of HANI (the ratio of hydraulic 377 

conductivity along columns to hydraulic conductivity along rows) always being 1 and the values of 378 

VKA (the ratio of horizontal to vertical hydraulic conductivity) fixed as the values in the original 379 

MODFLOW set-up. For MODFLOW parameters, the originally calibrated and modified parameter 380 

values in the steady-state MODFLOW version were used as the initial parameter values in PEST, and 381 

a small range around the initial values was assigned as the parameter range according to the experience 382 

from manual calibration and representative values (derived from http://www.aqtesolv.com/aquifer-383 

tests/aquifer_properties.htm). 384 

Table 3. 385 

The observed streamflow used for calibrating SWAT-MODFLOW was identical to that used for 386 

calibrating SWAT. Relatively continuous observations of the groundwater table were available at the 387 

location of two grid cells, and these were used for calibrating the variation of the groundwater table 388 

simulated by SWAT-MODFLOW. Because station A is located upstream from station B and its flow 389 

thus has an influence on station B, the weight for deriving the objective function for station A, which 390 

represents a small stream, was set to 2, and the weight for station B was set to 1. The weights for the 391 

two grid cells were set to 1. 392 

In order to establish template files and facilitate the process of modifying parameter values (HK, SS, 393 

SY) in the UPW Package while running PEST, the parameter value file (PVAL) and Zone file 394 

Page 14: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

14  

(https://water.usgs.gov/ogw/modflow-nwt/MODFLOW-NWT-Guide/) were first established based on 395 

the original UPW Package through running a code file in FORTRAN. 396 

Ten iterations were specified as the stop criteria in the PEST control file. Due to the large number of 397 

grid cells (183,112) in the MODFLOW set-up and the amount of disaggregated HRUs (DHRUs, 398 

66,765) compared with the case study conducted by (Bailey et al., 2017), it takes the coupled SWAT-399 

MODFLOW model complex around 4 hours to run a single simulation (12 years’ simulation) when 400 

MODFLOW runs with a daily interval. To shorten the calibration time, 11 BEOPEST slaves were 401 

created on three computers with BEOPEST as the pest executable file so that 11 simulations could be 402 

run simultaneously. A total of 638 simulations were run before the stop criteria was achieved. With 403 

the calibrated parameters fixed, the water stress threshold was calibrated manually to ensure proper 404 

simulation of the annual irrigation amount (ranging from 80 to 120 mm yr-1, occurring in the period 405 

between April to October) and make the simulated average annual irrigation amount in the irrigated 406 

HRUs (mm yr-1) comparative with that in the calibrated SWAT model. Finally, the SWAT-407 

MODFLOW model performance was validated following a procedure equivalent to that used for 408 

SWAT. 409 

2.4 Water abstraction scenarios 410 

Besides the 158 wells registered within the Uggerby River Catchment, another 256 wells exist outside 411 

the catchment but inside Hjørring Municipality (Fig. 3). All these wells were included in the Well 412 

Package in the SWAT-MODFLOW set-up. In SWAT-MODFLOW, the irrigation pumping source was 413 

defined as the third layer. For drinking water wells, 7 of the 101 drinking water wells were placed in 414 

the first layer, 91 in the third layer and 3 in the fifth layer. In order to evaluate the impacts of both 415 

irrigation and drinking water abstractions on streamflow for streams of difference sizes, four 416 

abstraction scenarios were designed and applied to the Uggerby River Catchment using both models: 417 

1) the no-wells scenario, where all abstractions are terminated; 2) the irrigation-wells-stop scenario, 418 

where all abstractions in irrigation wells are terminated, while abstractions in drinking water wells 419 

remain; 3) the drinking-wells-stop scenario, where all abstractions in drinking water wells are 420 

terminated, while abstractions in irrigation wells remain; and 4) the baseline scenario, where 421 

abstractions in all wells are included, which represents the current level of abstraction. We assumed 422 

that the point source discharge to the stream in subbasin 16 would remain the same in all scenarios. 423 

Once the scenarios were simulated, their impacts on streamflow were analyzed by assessing the 424 

average annual runoff amount, the contribution of water balance components, and the temporal 425 

Page 15: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

15  

dynamics of streamflow. The simulated signals of SWAT and SWAT-MODFLOW in the abstraction 426 

scenarios were then compared. 427 

3 Results 428 

3.1 Steady-state MODFLOW performance 429 

Visualization of the proximity of simulated and observed head contours (Fig. 6) was used to evaluate 430 

how well the modified calibrated MODFLOW model performed at steady state and three summary 431 

statistics were used as indicators for goodness of model fit (Table 4). The simulated heads and 432 

summary statistics have changed little compared with the original calibrated MODFLOW set-up. Thus, 433 

the modified calibrated MODFLOW model was satisfactory and suitable as a basis for coupling to 434 

SWAT in transient mode. 435 

Figure 6. 436 

Table 4. 437 

3.2 SWAT and SWAT-MODFLOW transient model performance 438 

The SWAT and SWAT-MODFLOW models both represented well the streamflow hydrographs during 439 

the calibration period, while during the validation period, one high peak flow event occurred in the 440 

SWAT and SWAT-MODFLOW simulations but not in the observations (Fig. 7). The baseflow was 441 

generally reproduced well by both models, but the SWAT-MODFLOW visibly performed better. 442 

Figure 7. 443 

Table 5. 444 

Compared with the recommended evaluation criteria by  (Moriasi et al., 2015), the statistical 445 

performance (Table 5) suggested “very good” performance of both models during the calibration 446 

period based on percent bias (PBIAS). During the validation period, the models performed “good” at 447 

station A and “satisfactory” at station B. For NSE values, the performance was “very good” for SWAT-448 

MODFLOW calibration at station B, “good” for SWAT-MODFLOW calibration at station A, 449 

“satisfactory” for SWAT calibration and SWAT-MODFLOW validation at both stations and SWAT 450 

validation at station A, but “unsatisfactory” for SWAT validation at station B. For R2 values, the 451 

Page 16: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

16  

performance was “good” for SWAT-MODFLOW calibration, “satisfactory” for SWAT calibration 452 

and SWAT-MODFLOW validation, but “unsatisfactory” for SWAT validation. 453 

The statistical performances of SWAT-MODFLOW with and without PEST calibration were 454 

compared. After calibration by PEST, the summary statistics of SWAT-MODFLOW were improved, 455 

especially for the validation period at station B where the performance increased from “unsatisfactory” 456 

to “satisfactory” according to NSE values (Table 5). In addition, the weighted residuals between 457 

simulation and observation were reduced after calibration by PEST, with the reduced residuals mainly 458 

coming from streamflow simulation (Table 6). 459 

Table 6. 460 

In SWAT, almost all the top 12 sensitive parameters (Fig. 8) were surface process parameters (Table 461 

2) except for the groundwater parameter GW_DELAY. In contrast, for SWAT-MODFLOW (Table 3), 462 

all the top 12 sensitive parameters were groundwater parameters with the exclusion of only one surface 463 

process parameter OV_N. 464 

Figure 8. 465 

Compared with SWAT, the SWAT-MODFLOW model not only produced output for streamflow but 466 

also for the groundwater table of each cell on any given day. The variation of groundwater heads across 467 

the simulation period was minimal for layer 1, while there was some, albeit small, variation in layer 3 468 

(Fig. 9). There was generally a good agreement between the groundwater head level and dynamics 469 

simulated by SWAT-MODFLOW and that was recorded at the two observation wells within the 470 

catchment (Fig. 10). 471 

Figure 9. 472 

Figure 10. 473 

For the water balance, the evaporation simulated by SWAT-MODFLOW was a little higher (13 mm 474 

yr-1) than that simulated by SWAT, while the total water yields (total stream flow) simulated by SWAT 475 

and SWAT-MODFLOW were almost equal (Table 7). The water balance components, however, 476 

differed substantially. Compared with SWAT, the surface runoff simulated by SWAT-MODFLOW 477 

was a little higher, while the lateral subsurface flow and groundwater flow (simulated by the River 478 

Package) were much lower. In SWAT-MODFLOW, the largest contributor to streamflow was the 479 

Page 17: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

17  

drain flow simulated by the Drain Package (constituting 70% of the streamflow). Conceptually, 480 

however, this can also be viewed as a surface-near groundwater contribution. Hence, when lumping 481 

the contribution from drains and groundwater, these are clearly the dominant sources for streamflow 482 

in both the SWAT and SWAT-MODFLOW model. 483 

Table 7. 484 

3.3 Water abstraction scenarios simulation 485 

The annual abstrations by drinking water wells or irrigation wells set up in the two models were 486 

approximately equivalent (Table 8). In the SWAT simulations, compared with the no-wells scenario 487 

(scenario 1), a decrease in the average annual stream flow was observed in scenario 2 (only drinking 488 

water wells), while an increase was recorded in scenario 3 (only irrigation wells) and scenario 4 (both 489 

drinking water and irrigation wells). In the SWAT-MODFLOW simulations, the average annual 490 

streamflow decreased not only in scenario 2, but also in scenario 4, and at subbasin 18 outlet in 491 

scenario 3, while a slight increase occurred at subbasin 13 outlet in scenario 3. The decrease in scenario 492 

2 simulated by SWAT-MODFLOW was much larger than that by SWAT and also closer to the 493 

abstracted amount, and the increase at subbasin 13 outlet in scenario 3 simulated by SWAT-494 

MODFLOW was apparently lower than that simulated by SWAT (Table 8). 495 

Table 8. 496 

In SWAT, the decrease of average annual total flow in scenario 2 was minimal as a result of a tiny 497 

decrease in the groundwater return flow (Fig. 11a). In scenario 3 and scenario 4, with unchanged tile 498 

flow, all the other flow components rose, especially groundwater and lateral soil discharge. In SWAT-499 

MODFLOW, the decrease of average annual total flow in scenario 2 also resulted from a decreased 500 

groundwater return flow, but the decrease was much larger than that simulated by SWAT. In scenario 501 

3, the lateral soil runoff and drain flow increased in SWAT-MODFLOW similar to SWAT, while in 502 

scenario 4, reduced drain flow was recorded (Fig. 11b). Compared with the no-wells scenario, the 503 

amount of evapotransportation remained unchanged in scenario 2, whereas it increased by 5 mm yr-1 504 

in the sceanarios with irrigation wells in both the SWAT and SWAT-MODFLOW simulations. In the 505 

scenario with only irrigation, evaportransportation and total flow increased in both the SWAT and 506 

SWAT-MODFLOW simulations, but the soil or aquifer water storage decreased according to the water 507 

balance. 508 

Page 18: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

18  

Figure 11. 509 

When comparing the temporal patterns of streamflow with the no-wells scenario (scenario 1), we found 510 

the daily discharge difference in scenario 2 (only drinking water wells) to be almost always negative 511 

(sometimes zero), while in scenario 3 (only irrigation wells) and scenario 4 (both drinking water and 512 

irrigation wells) it fluctuated around zero in simulations by both SWAT and SWAT-MODFLOW (Fig. 513 

12). Thus, the daily flow in the scenario with drinking water wells was almost always lower than the 514 

scenario without drinking water wells, and the daily flow in the scenario with only irrigation wells or 515 

the scenario with both irrigation and drinking water wells could be higher or lower than the scenario 516 

without wells. The daily discharge difference between scenario 2 and the no-wells scenario simulated 517 

by SWAT-MODFLOW was obvious, but the difference using SWAT was minimal. In the comparison 518 

of scenario 3 with the no-wells scenario, when the discharge difference was positive after an irrigation 519 

event, it descended smoothly in the SWAT simulation and more sharply in the SWAT-MODFLOW 520 

simulations. 521 

Figure 12. 522 

In the SWAT-MODFLOW set-up, the water exchange between aquifer and streams occurs between 523 

each MODFLOW river/drain cell and its surrounding cells. The newly developed SWAT-MODFLOW 524 

model complex can output the daily rate of water exchange between aquifer and streams for each 525 

subbasin. When the water exchange is positive, it is indicative of water flow from the aquifer to the 526 

stream. The temporal pattern of groundwater discharge was the same as for the stream flow, and the 527 

temporal patterns of the differences in groundwater discharge between the abstraction scenarios and 528 

the no-wells scenario were similar to the differences in streamflow, except for some peak flow days 529 

(Fig. 13), which indicates that the abstraction-induced streamflow change followed the groundwater 530 

discharge change. 531 

Figure 13. 532 

4. Discussion 533 

4.1 Performance and parameter sensitivity of SWAT and SWAT-MODFLOW 534 

Both the SWAT and SWAT-MODFLOW model simulated the temporal patterns of streamflow 535 

generally well at the two hydrological stations during the calibration and validation periods. However, 536 

visually SWAT-MODFLOW performed better, especially during recession curves and low flow 537 

Page 19: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

19  

periods, suggesting a better simulation of the interaction between surface water and groundwater. 538 

Accordingly, the corresponding summary performance statistics were also better for SWAT-539 

MODFLOW. The simulated peak flow on 16 October 2014 by both models was much higher than the 540 

observed data (Fig. 7). This discrepancy may be attributed to a high record of precipitation on that day 541 

based on a 10 by 10 km grid, which may not be representative for the wider catchment. Additionally, 542 

it is also likely that the observed streamflow was underestimated as it is calculated from the Q-h 543 

relation, which typically does not adequately cover peak flow events (Poulsen, 2013). 544 

In the parameter sensitivity analysis, the surface process parameters of the two models shared the same 545 

ranges, while the models had different groundwater modules and parameters. While the SWAT-546 

MODFLOW calibration was based on an objective function that took into account not only streamflow 547 

but also groundwater heads at the location of two wells, the calibration by PEST mainly improved the 548 

streamflow simulation performance (Table 4). According to the parameter sensitivity ranking, the 549 

parameters regarding groundwater processes in SWAT-MODFLOW played an important role in the 550 

streamflow simulation performance, while in SWAT, the impact of groundwater module parameters 551 

on streamflow simulation was generally insignificant. This reflects the shortcoming of the SWAT 552 

groundwater module, which ignores the variability in distributed parameters such as hydraulic 553 

conductivity and storage coefficients, represents groundwater by a lumped module in individual 554 

subbasins, and contributes to the stream network as baseflow based on a linear reservoir approximation. 555 

With this simplified implementation of groundwater dynamics and water exchange between surface 556 

water and groundwater in SWAT, the discharge simulated by SWAT cannot be optimized to the same 557 

extent as that simulated by SWAT-MODFLOW. 558 

The availability of spatial-temporal patterns of the groundwater head in SWAT-MODFLOW could 559 

significantly benefit groundwater resources management and provide yet another level of 560 

understanding of water resources dynamics within a catchment. The outputs of SWAT-MODFLOW 561 

in this study showed that the model performed well, not only in streamflow simulations but also with 562 

respect to the spatial-temporal patterns of the simulated groundwater head. In contrast, since no 563 

information of groundwater table output is provided by SWAT, its goodness in streamflow simulation 564 

may potentially be based on an improper groundwater simulation where its performance on 565 

groundwater simulation is unknown. 566 

4.2 Models ability to simulate effects of groundwater abstractions on streamflow 567 

Page 20: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

20  

In scenario 2 where only drinking water wells are active according to the water balance where there is 568 

no change in evaporation compared with the no-wells scenario, we expected that the streamflow 569 

depletion simulated by SWAT would be approximately equivalent to the abstracted water volume, 570 

taking into account a possible small change in the aquifer or soil storage. However, results in this study 571 

showed that the impact of drinking water abstractions on streamflow in the SWAT simulation was 572 

negligible. In the SWAT-MODFLOW set-up, because the aquifer in the Uggerby River Catchment is 573 

connected to and interactive with an area outside of the topographical catchment (Fig. 3), the 574 

abstraction from an aquifer located in the Uggerby River Catchment not only impacts the hydrology 575 

inside but potentially also outside the catchment. According to the water balance, we expected that the 576 

SWAT-MODFLOW simulated streamflow depletion in the catchment would be lower at a level 577 

somewhat equivalent to the abstracted water volume. With equivalent abstraction for drinking water, 578 

the annual flow decrease simulated by SWAT-MODFLOW was much larger than that by SWAT and 579 

closer to the abstracted volume. Therefore, we conclude that SWAT simulations underestimate the 580 

impacts of groundwater abstraction for drinking water on streamflow depletion, while SWAT-581 

MODFLOW provided more realistic assessments. 582 

The simulated irrigation operation abstracts water from an aquifer and then applies the water onto the 583 

surface of agricultural land or pasture. Most of the water infiltrates back into the soil and is then utilized 584 

by the vegetation and partly lost through evapotranspiration or infiltrates deeper to the aquifer, and a 585 

small part of the water might flow to streams directly as a small increase in surface runoff. Though the 586 

abstraction causes groundwater depletion, the recharge from the irrigated water can partly refill the 587 

aquifer and produce groundwater discharge. Since in the SWAT-MODFLOW set-up the aquifer in the 588 

Uggerby River Catchment was connected and interactive with an outside area, after each event of 589 

groundwater abstraction for irrigation, the aquifer storage would be recharged not only from the 590 

irrigated land area but also by the groundwater flowing from the outside area. If the recharge rate is 591 

larger than the abstracted water amount, the groundwater discharge to the stream will presumably 592 

increase. Hence, the irrigation events also brought about a slight increase of average annual stream 593 

flow at the subbasin 13 outlet (Table 8), and a slight total flow increase within the catchment (Fig. 594 

11b). The subbasin aquifers in the SWAT set-up are closed and have no interaction with areas outside 595 

a subbasin. Meanwhile, the abstracted amount of water from aquifers for irrigation is larger than the 596 

amount of returning aquifer recharge from irrigated water, and we would therefore expect decreasing 597 

groundwater discharge to streamflow in SWAT simulations. However, the SWAT simulations also 598 

showed that irrigation led to enhanced streamflow (Table 8, Fig. 11a), which apparently was even 599 

higher than the increase simulated by SWAT-MODFLOW. This supports the point mentioned above 600 

Page 21: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

21  

that SWAT underestimates the abstraction effect on streamflow depletion. SWAT simulations can, 601 

therefore, lead to incorrect assessments of the impacts of groundwater abstractions for irrigation on 602 

streamflow, while SWAT-MODFLOW provided more realistic assessments. 603 

Upon inspecting the SWAT source code, it appears that the groundwater discharge calculation 604 

equation used in SWAT does not take into account the impact of water abstraction from shallow 605 

aquifers on water table fluctuations. Thus, the groundwater removal by abstractions in the SWAT 606 

simulation does not have a direct effect on the groundwater discharge, which may explain the 607 

somewhat surprising simulation signals of SWAT. In addition, in the equation, the groundwater 608 

discharge on the current day is highly related to the groundwater discharge on the previous day, and 609 

the increase of the groundwater discharge resulting from each irrigation application could then lead to 610 

enhanced groundwater discharge for several days in a row. This may explain why the increased 611 

discharge following an irrigation event descended more smoothly in SWAT than in SWAT-612 

MODFLOW (Fig. 12). 613 

In the SWAT-MODFLOW model, the exchange rate between groundwater and surface water is based 614 

on the head difference between the river stage (or drain cell stage) and the head of its surrounding 615 

groundwater grid cells. This can reflect the temporally dynamic hydrological processes and also the 616 

impacts from all the external stressors (e.g. temporally and spatially varying recharge and groundwater 617 

abstractions) on water table fluctuations. Naturally, this should also allow SWAT-MODFLOW to 618 

provide more realistic assessments of the impacts of groundwater abstractions on streamflow in 619 

comparison with SWAT. 620 

While setting up the drinking water abstraction in SWAT, three limitations were identified, also 621 

reported in (Molina-Navarro et al., 2019). The first is that SWAT only allows one decimal point for 622 

abstraction numerical inputs with a unit of 104 m3 day-1 for each month. This means that pumping rate 623 

variations within one month cannot be simulated by SWAT and that the accuracy of abstraction 624 

dynamics thus cannot be guaranteed. As a result of this limitation, the abstraction amount in SWAT 625 

and SWAT-MODFLOW was not completely identical. The second limitation is that the abstraction 626 

from deep aquifer did not result in any streamflow change. Therefore, all the abstraction sources had 627 

to be defined as the shallow aquifer in SWAT to achieve a signal in streamflow despite that we had at 628 

least three wells receiving water from a deep aquifer (the fifth layer according to the MODFLOW-629 

NWT set-up). The last limitation is that the abstraction rates of all wells in each subbasin in SWAT 630 

have to be summed up to one input value, thereby ignoring the specific location of wells within 631 

individual subbasins. 632 

Page 22: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

22  

SWAT-MODFLOW overcomes the limitations in SWAT by exploiting the spatial explicitness of 633 

MODFLOW where groundwater abstraction can be simulated using the Well Package, which allows 634 

many decimal points for abstraction inputs as well as user-defined units, pumping rates at potentially 635 

daily intervals, and wells located in any vertical layer and any grid cell within a subbasin. In addition 636 

to the outputs from SWAT, SWAT-MODFLOW also provides fully distributed groundwater-related 637 

outputs such as spatial-temporal patterns of water table elevation, distributed aquifer recharge, and 638 

groundwater-surface water exchange rates at a cell level, permitting detailed analysis of groundwater 639 

and its interaction with surface water. This may be an important input to groundwater resources 640 

management (e.g. groundwater abstraction) and the solving of surface water rights issues. These 641 

capabilities demonstrate the advantage of SWAT-MODFLOW over modifying the SWAT 642 

groundwater module codes to improve groundwater flow simulation (Nguyen and Dietrich, 2018; 643 

Pfannerstill et al., 2014; Zhang et al., 2016), which remains a semi-distributed way to simulate 644 

subsurface hydrologic processes and does not generate detailed groundwater outputs. This point 645 

supports the findings about the advantages of SWAT-MODFLOW over SWAT in (Molina-Navarro et 646 

al., 2019) but using a much more complex set-up. 647 

4.3 Performance of SWAT-MODFLOW and SWAT relative to other recent studies 648 

In previous studies, after coupling a calibrated SWAT and calibrated MODFLOW model, the SWAT-649 

MODFLOW model complex was applied without further calibration (Bailey et al., 2016; Chunn et al., 650 

2019), with calibration against only streamflow  observations  (Molina-Navarro et al., 2019), with 651 

separated calibration for streamflow and groundwater head  (Guzman et al., 2015), or with simple 652 

manual calibration by graphically comparing the simulated and observed streamflow and groundwater 653 

head (Sith et al., 2019). Since both the SWAT and MODFLOW supporting software can use the inverse 654 

modeling (IM) method for calibration, and parameter non-uniqueness is an inherent property of IM 655 

(Abbaspour, 2015), the coupling of a calibrated SWAT and a calibrated MODFLOW cannot guarantee 656 

a proper or sufficiently optimized parameter set for the integrated SWAT-MODFLOW model. Because 657 

groundwater and surface water interact with each other, calibrating the simulation of one part does not 658 

guarantee proper simulation of the other part. Application of a combined calibration approach based 659 

on PEST allowed us to calibrate the SWAT-MODFLOW model by adjusting simultaneously SWAT 660 

and MODFLOW parameters and using observations of both streamflow and groundwater table when 661 

deriving the objective function. The calibration results demonstrated that the summary statistics of the 662 

SWAT-MODFLOW performance were improved by this approach (Table 6). 663 

Page 23: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

23  

The ability of SWAT-MODFLOW to evaluate the impacts of groundwater abstraction on streamflow 664 

or groundwater-surface water interactions has been tested in the previous studies (Guzman et al., 2015; 665 

Chunn et al., 2019; Molina-Navarro et al., 2019). (Molina-Navarro et al., 2019), for example, also 666 

found that the SWAT model showed almost no impact of groundwater abstraction on streamflow 667 

depletion. Besides due to the simple representation of groundwater dynamics, the other cause of this, 668 

we believe, is that same as suggested above, that the impact of groundwater water removal by 669 

abstractions on water table fluctuations is currently not accounted for in the groundwater discharge 670 

calculation in the SWAT source code. Our findings are generally consistent with those of these 671 

previous studies, although all of the studies tested the effects of groundwater abstraction only by 672 

drinking water without considering irrigation and based on assumed drinking water pumping wells. In 673 

addition, in all the previous SWAT-MODFLOW studies, the River Package in the MODFLOW model 674 

was the only package used for simulating groundwater-surface water interaction, ignoring the potential 675 

drain flow processes. The SWAT-MODFLOW complex used in our study was further developed to 676 

allow application of the Drain Package and to allow also an auto-irrigation routine to extract water 677 

from groundwater grid cells; in this way the impacts of groundwater abstraction for both drinking 678 

water and irrigation could be assessed. 679 

4.4 Limitations and future research 680 

Several limitations to this study need to be acknowledged. The simulated head generated by the steady-681 

state model was used as the initial head conditions for the transient model, as also suggested in other 682 

studies (Anderson et al., 2015; Doherty et al., 2010). The ideal simulated initial heads should be 683 

calibrated with the observed initial heads. However, we did not have enough observed heads at the 684 

beginning of the simulation period (1997), so we used the observed heads covering the period 1996-685 

2010 for calibrating the original steady-state MODFLOW-NWT to obtain the simulated initial heads. 686 

Fortunately, the groundwater heads of the study area did not change much during the study period (Fig. 687 

9, Fig. 10) and the difference inherently exists between the observed and simulated heads, indicating 688 

that the error between the ideal simulated initial heads and the actually used simulated initial heads 689 

was small. 690 

An approach based on PEST was utilized to calibrate streamflow and groundwater table variation 691 

simultaneously in our SWAT-MODFLOW simulation, which improved the model performance and 692 

enabled parameter sensitivity analysis for the model. However, only two wells with relatively 693 

continuous time series of observed groundwater head were available and used to calibrate the 694 

groundwater variation. Ideally, calibration would involve more wells with continuous time series of 695 

Page 24: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

24  

observed head, but this limitation is anticipated to be minor in our study as the groundwater head did 696 

not change much in our simulations and the change mainly followed the variation of recharge with 697 

precipitation as its source. 698 

The average annual streamflow difference and the regular pattern of daily streamflow difference 699 

between the abstraction scenarios and the no-wells scenario were generally explained well, but, 700 

surprisingly and unexpectedly, the streamflow difference between the scenario with only drinking 701 

water wells and the no-wells scenario on 24 March, 2010, simulated by SWAT-MODFLOW at two 702 

stations, were positive, being 1.54 and 0.55 m3 s-1, respectively (Fig. 12). The streamflow difference 703 

between the scenario with only irrigation wells and the no-wells scenario at station B on the extreme 704 

peak flow day (16 October, 2014) simulated by SWAT was -5.2 m3 s-1 but then became positive next 705 

day, which cannot be explained well to our best of knowledge so far. However, we found that the 706 

general results of this study were not influenced when modifying the value of these two unexpected 707 

points to be expected. 708 

Both the SWAT and SWAT-MODFLOW simulations were based on the “best” parameter combination 709 

achieved through calibration, which was deemed to be satisfactory for the purpose of this study. 710 

However, complex models such as SWAT and SWAT-MODFLOW are subject to non-uniqueness (i.e. 711 

more than one parameter combination may yield satisfactory results), so future studies may need to 712 

consider the uncertainty due to, for example, parameter uncertainty. The calibration tool SWAT-CUP 713 

has already been able to evaluate SWAT parameter uncertainty, whereas the new approach based on 714 

PEST to calibrate SWAT-MODFLOW needs to be further explored to adapt for model uncertainty 715 

analysis. 716 

Our results support our original hypothesis that SWAT-MODFLOW can produce more reliable results 717 

in the simulation of the effects of groundwater abstraction for either drinking water or irrigation on 718 

streamflow patterns. In addition, SWAT-MODFLOW can produce more outputs than SWAT. 719 

However, SWAT-MODFLOW also requires more effort and data to be set up and calibrated, and 720 

longer time to run (around 6 hours for a 19-year simulation in SWAT-MODFLOW by a desktop with 721 

an Intel® Core™ Processor i7-6700 CPU and 16 GB installed RAM versus 6 minutes for a SWAT 722 

simulation). Therefore, the balance between scientific accuracy and the computational burden should 723 

be defined relative to the study goal when choosing between SWAT and SWAT-MODFLOW in a 724 

future study. But clearly, if the purpose of a study is to investigate effects of groundwater abstraction 725 

on streams, the efforts should be focused on setting up and applying a fully-distributed model in 726 

groundwater domain, such as SWAT-MODFLOW. A graphical user interface has also been developed 727 

Page 25: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

25  

to couple SWAT and MODFLOW based on the publically available version of the SWAT-728 

MODFLOW complex (Park et al., 2018). Since the SWAT-MODFLOW complex used in this study 729 

was newly developed and allowed use of the Drain Package and auto-irrigation, a new graphical user 730 

interface based on the new SWAT-MODFLOW complex could ensure that a study such as that 731 

presented here is repeated with less effort and technical challenges. 732 

5. Conclusions 733 

SWAT and SWAT-MODFLOW models with relatively complex set-ups were applied to a lowland 734 

catchment, the Uggerby River Catchment in Northern Denmark. Model performance and the outcome 735 

of four groundwater abstraction scenarios (with real wells and abstraction rates) were analyzed and 736 

compared. 737 

Generally both models simulated well the temporal patterns of streamflow at the two hydrological 738 

stations during the calibration and validation periods. SWAT-MODFLOW, however, showed superior 739 

performance when visualizing time series results and when comparing summary statistics. 740 

Furthermore, SWAT-MODFLOW generates many additional outputs for groundwater analysis, such 741 

as spatial-temporal patterns of water table elevation and groundwater-surface water exchange rates at 742 

cell or subbasin level, improving water resources management in a groundwater-dominated catchment. 743 

Abstraction scenarios simulated by SWAT and SWAT-MODFLOW showed different signals in 744 

streamflow change. The simulations by both models indicated that drinking water abstraction caused 745 

streamflow depletion and that irrigation abstraction caused a slight total flow increase (but decreased 746 

the soil or aquifer water storage, which may influence the hydrology outside the catchment). However, 747 

the impact of drinking water abstraction on streamflow depletion by SWAT was minimal and 748 

underestimated, and the streamflow increase caused by irrigation abstraction was  exaggerated 749 

compared with the SWAT-MODFLOW simulation, which produced more realistic results. 750 

Overall, the new SWAT-MODFLOW model calibrated by PEST, which included the Drain Package 751 

and a new auto-irrigation routine, presented a better hydrological simulation, wider possibilities for 752 

groundwater analysis, and more realistic assessments of the impact of groundwater abstractions (for 753 

either irrigation or drinking water purposes) on streamflow compared with SWAT. Thus, SWAT-754 

MODFLOW can be used as a tool for managing water resources in groundwater-affected catchments, 755 

taking into account its higher computational demand and more time consumption. 756 

Page 26: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

26  

Code and data availability. 757 

The land use map based on the Danish Area Information System is freely available from 758 

(https://www.dmu.dk/1_viden/2_miljoe-tilstand/3_samfund/ais/3_Metadata/metadata_en.htm). 759 

Climate data is available from the Danish Meteorological Institute (https://www.dmi.dk/). QGIS is 760 

freely available from https://qgis.org/en/site/. QSWAT, SWATCUP, and the SWAT-MODFLOW as 761 

well as its source codes are publicly available from https://swat.tamu.edu/software. The steady-state 762 

MODFLOW set-up was provided by NIRAS upon request. The PEST utilities and tutorial are freely 763 

downloadable from http://www.pesthomepage.org/Home.php. The source code, executable, and tutorial 764 

for the further developed SWAT-MODFLOW are available on the SWAT website 765 

(https://swat.tamu.edu/software/swat-modflow/). The two code files used for SWAT-MODFLOW 766 

calibration by PEST will be available through repository on https://www.re3data.org/ when the paper 767 

is accepted. 768 

769 

Author contributions. DT and WL designed the study. WL undertook all practical elements of the 770 

study, including setting up and calibrating the models, analyzing results, producing figures, and writing 771 

the manuscript. SP and RTB contributed the idea and developed the codes for use of the PEST 772 

approach to calibrate the SWAT-MODFLOW model. RTB provided the knowledge to set up SWAT-773 

MODFLOW and further developed the SWAT-MODFLOW complex codes. DT, EMN, HEA, HT, 774 

and AN provided most of the data and contributed with their knowledge to setting up and calibrating 775 

the models. DT and EJ helped to analyze the results and contributed to the discussion. JSJ and JBJ 776 

provided the original steady-state MODFLOW-NWT set-up and contributed with relative knowledge. 777 

All co-authors contributed to the manuscript writing. 778 

779 

Competing interests. The authors declare that they have no conflict of interest. 780 

781 

Acknowledgements. 782 

The first author was supported by grants from the China Scholarship Council. Erik Jeppesen and 783 

Dennis Trolle were supported by the AU Centre for Water Technology (WATEC). We thank Chenda 784 

Deng, Xiaolu Wei, and Zaichen Xiang for technique assistance and knowledge exchange during Wei 785 

Liu’s research stay at the Colorado State University. We also thank Anne Mette Poulsen for valuable 786 

editorial comments. 787 

788 

789 

790 

Page 27: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

27  

791 

References 792  793 Abbas, S., Xuan, Y., and Bailey, R.:  Improving River Flow Simulation Using a Coupled Surface‐Groundwater 794 Model for Integrated Water Resources Management, 2018. 795 Abbaspour, K. C.: SWAT‐CUP: SWAT Calibration and Uncertainty Programs ‐ A User Manual., 2015. 796 Ali, R., McFarlane, D., Varma, S., Dawes, W., Emelyanova, I., Hodgson, G., and Charles, S.: Potential climate 797 change impacts on groundwater resources of south‐western Australia, Journal of Hydrology, 475, 456‐472, 798 https://doi.org/10.1016/j.jhydrol.2012.04.043, 2012. 799 Aliyari, F., Bailey, R. T., Tasdighi, A., Dozier, A., Arabi, M., and Zeiler, K.: Coupled SWAT‐MODFLOW model for 800 large‐scale  mixed  agro‐urban  river  basins,  Environmental  Modelling  &  Software,  115,  200‐210, 801 https://doi.org/10.1016/j.envsoft.2019.02.014, 2019. 802 Anderson, M. P., Woessner, W. W., and Hunt, R. J.: Chapter 7 ‐ Steady‐State and Transient Simulations, in: 803 Applied Groundwater Modeling (Second Edition), edited by: Anderson, M. P., Woessner, W. W., and Hunt, R. 804 J., Academic Press, San Diego, 303‐327, 2015. 805 Aslyng, H.: Forelæsninger over vanding i jordbruget, DSR Forlag. Den kgl. Veterinær‐og Landbohøjskole, 1983. 806 Bailey, R., Rathjens, H., Bieger, K., Chaubey,  I., and Arnold,  J.:  Swatmod‐Prep: Graphical User  Interface  for 807 Preparing Coupled Swat‐Modflow Simulations, J Am Water Resour As, 53, 400‐410, 10.1111/1752‐1688.12502, 808 2017. 809 Bailey, R. T., Wible, T. C., Arabi, M., Records, R. M., and Ditty,  J.: Assessing  regional‐scale  spatio‐temporal 810 patterns of groundwater‐surface water  interactions using a coupled SWAT‐MODFLOW model, Hydrological 811 Processes, 30, 4420‐4433, 10.1002/hyp.10933, 2016. 812 Cheema, M. J., Immerzeel, W. W., and Bastiaanssen, W. G.: Spatial quantification of groundwater abstraction 813 in the irrigated Indus basin, Ground Water, 52, 25‐36, 10.1111/gwat.12027, 2014. 814 Chen, X. H., and Yin, Y. F.: Streamflow depletion: Modeling of reduced baseflow and induced stream infiltration 815 from seasonally pumped wells, J Am Water Resour As, 37, 185‐195, DOI 10.1111/j.1752‐1688.2001.tb05485.x, 816 2001. 817 Chen, Y., Marek, G., Marek, T., Brauer, D., and Srinivasan, R.: Assessing the Efficacy of the SWAT Auto‐Irrigation 818 Function  to  Simulate  Irrigation,  Evapotranspiration,  and  Crop  Response  to Management  Strategies  of  the 819 Texas High Plains, Water‐Sui, 9, 10.3390/w9070509, 2017. 820 Chunn, D., Faramarzi, M., Smerdon, B., and Alessi, D.: Application of an Integrated SWAT–MODFLOW Model 821 to  Evaluate  Potential  Impacts  of  Climate Change  and Water Withdrawals  on Groundwater–Surface Water 822 Interactions in West‐Central Alberta, Water‐Sui, 11, 10.3390/w11010110, 2019. 823 Doherty, J.: PEST: Model‐independent parameter estimation and Uncertainty Analysis, User manual: [EB/OL], 824 Brisbane, Queensland, Australia: Watermark Numeric Computing, 2018. 825 Doherty, J. E., Hunt, R. J., and Tonkin, M. J.: Approaches to highly parameterized inversion: A guide to using 826 PEST for model‐parameter and predictive‐uncertainty analysis, US Geological Survey Scientific Investigations 827 Report, 5211, 71, 2010. 828 Fukunaga, D. C., Cecílio, R. A.,  Zanetti,  S.  S., Oliveira,  L.  T.,  and Caiado, M. A. C.: Application of  the SWAT 829 hydrologic model to a tropical watershed at Brazil, Catena, 125, 206‐213, 10.1016/j.catena.2014.10.032, 2015. 830 Gao, F., Feng, G., Han, M., Dash, P., Jenkins, J., and Liu, C.: Assessment of Surface Water Resources in the Big 831 Sunflower River Watershed Using Coupled SWAT–MODFLOW Model, Water‐Sui, 11, 528, 2019. 832 Gassman, P. W., Sadeghi, A. M., and Srinivasan, R.: Applications of the SWAT Model Special Section: Overview 833 and Insights, J Environ Qual, 43, 1‐8, 10.2134/jeq2013.11.0466, 2014. 834 George, Y. D. R. S. C.: QGIS Interface for SWAT (QSWAT), 2017. 835 GEUS: Water supply in Denmark, Danish Ministry of the Environment, Denmark, 18, 2009. 836 Glover, R. E., and Balmer, G. G.: River depletion resulting from pumping a well near a river, Eos, Transactions 837 American Geophysical Union, 35, 468‐470, 1954. 838 Greve, M. H., Greve, M. B., Bøcher, P. K., Balstrøm, T., Breuning‐Madsen, H., and Krogh, L.: Generating a Danish 839 raster‐based topsoil property map combining choropleth maps and point information, Geografisk Tidsskrift‐840 Danish Journal of Geography, 107, 1‐12, 10.1080/00167223.2007.10649565, 2007. 841 

Page 28: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

28  

Güngör, Ö.,  and Göncü, S.: Application of  the  soil  and water assessment  tool model on  the Lower Porsuk 842 Stream Watershed, Hydrological Processes, 27, 453‐466, 10.1002/hyp.9228, 2013. 843 Guzman, J. A., Moriasi, D. N., Gowda, P. H., Steiner, J. L., Starks, P. J., Arnold, J. G., and Srinivasan, R.: A model 844 integration framework for linking SWAT and MODFLOW, Environmental Modelling & Software, 73, 103‐116, 845 10.1016/j.envsoft.2015.08.011, 2015. 846 Henriksen, H.  J., Troldborg, L., Højberg, A. L., and Refsgaard,  J. C.: Assessment of exploitable groundwater 847 resources of Denmark by use of ensemble resource indicators and a numerical groundwater–surface water 848 model, Journal of Hydrology, 348, 224‐240, 10.1016/j.jhydrol.2007.09.056, 2008. 849 Huang, C.‐S., Yang, T., and Yeh, H.‐D.: Review of analytical models to stream depletion induced by pumping: 850 Guide to model selection, Journal of Hydrology, 561, 277‐285, https://doi.org/10.1016/j.jhydrol.2018.04.015, 851 2018. 852 Hunt, B.: Unsteady stream depletion from ground water pumping, Groundwater, 37, 98‐102, 1999. 853 Izady,  A.,  Davary,  K.,  Alizadeh,  A.,  Ziaei,  A.  N.,  Akhavan,  S.,  Alipoor,  A.,  Joodavi,  A.,  and  Brusseau, M.  L.: 854 Groundwater  conceptualization  and  modeling  using  distributed  SWAT‐based  recharge  for  the  semi‐arid 855 agricultural Neishaboor plain, Iran, Hydrogeol J, 23, 47‐68, 10.1007/s10040‐014‐1219‐9, 2015. 856 Jeppesen, E., Brucet, S., Naselli‐Flores, L., Papastergiadou, E., Stefanidis, K., Nõges, T., Nõges, P., Attayde, J. L., 857 Zohary, T., Coppens, J., Bucak, T., Menezes, R. F., Freitas, F. R. S., Kernan, M., Søndergaard, M., and Beklioğlu, 858 M.: Ecological  impacts of global warming and water abstraction on  lakes and reservoirs due to changes  in 859 water level and related changes in salinity, Hydrobiologia, 750, 201‐227, 10.1007/s10750‐014‐2169‐x, 2015. 860 Johansen, O. M., Pedersen, M. L., and  Jensen,  J. B.: Effect of groundwater abstraction on  fen ecosystems, 861 Journal of Hydrology, 402, 357‐366, 10.1016/j.jhydrol.2011.03.031, 2011. 862 Kim, N. W., Chung, I. M., Won, Y. S., and Arnold, J. G.: Development and application of the integrated SWAT–863 MODFLOW model, Journal of Hydrology, 356, 1‐16, 10.1016/j.jhydrol.2008.02.024, 2008. 864 Knudsen, T., and Olsen, B. P.: Proceedings of the 2nd NKG workdshop on national DEMs, Technical report No.4, 865 National Survey and Cadastre, Danish Ministry of the Environment, Copenhagen, Denmark, 36, 2008. 866 Lachaal, F., Mlayah, A., Bédir, M., Tarhouni, J., and Leduc, C.: Implementation of a 3‐D groundwater flow model 867 in a semi‐arid region using MODFLOW and GIS tools: The Zéramdine–Béni Hassen Miocene aquifer system 868 (east‐central Tunisia), Computers & Geosciences, 48, 187‐198, 10.1016/j.cageo.2012.05.007, 2012. 869 Lee, K.‐S., Chung, E.‐S., and Shin, M.‐J.: Effects of changes of climate, groundwater withdrawal, and landuse 870 on total flow during dry period, Journal of Korea Water Resources Association, 39, 923‐934, 2006. 871 Liu,  W.,  An,  W.,  Jeppesen,  E.,  Ma,  J.,  Yang,  M.,  and  Trolle,  D.:  Modelling  the  fate  and  transport  of 872 Cryptosporidium, a zoonotic and waterborne pathogen, in the Daning River watershed of the Three Gorges 873 Reservoir  Region,  China,  J  Environ Manage,  232,  462‐474,  https://doi.org/10.1016/j.jenvman.2018.10.064, 874 2019. 875 Lu,  S.,  Kronvang,  B.,  Audet,  J.,  Trolle,  D.,  Andersen,  H.  E.,  Thodsen,  H.,  and  van  Griensven,  A.: Modelling 876 sediment  and  total  phosphorus export  from a  lowland  catchment:  Comparing  sediment  routing methods, 877 Hydrological processes, 29, 280‐294, 2015. 878 Lu, S., Andersen, H. E., Thodsen, H., Rubæk, G. H., and Trolle, D.: Extended SWAT model for dissolved reactive 879 phosphorus  transport  in  tile‐drained  fields  and  catchments,  Agricultural Water Management,  175,  78‐90, 880 10.1016/j.agwat.2015.12.008, 2016. 881 Malago, A., Bouraoui, F., Vigiak, O., Grizzetti, B., and Pastori, M.: Modelling water and nutrient fluxes in the 882 Danube River Basin with SWAT, Sci Total Environ, 603‐604, 196‐218, 10.1016/j.scitotenv.2017.05.242, 2017. 883 Markstrom, S. L., Niswonger, R. G., Regan, R. S., Prudic, D. E., and Barlow, P. M.: GSFLOW‐Coupled Ground‐884 water and Surface‐water FLOW model based on the integration of the Precipitation‐Runoff Modeling System 885 (PRMS) and the Modular Ground‐Water Flow Model (MODFLOW‐2005), US Geological Survey techniques and 886 methods, 6, 240, 2008. 887 Markstrom, S. L., Regan, R. S., Hay, L. E., Viger, R. J., Webb, R. M., Payn, R. A., and LaFontaine, J. H.: PRMS‐IV, 888 the precipitation‐runoff modeling system, version 4, US Geological Survey Techniques and Methods, 2015. 889 May,  R.,  and  Mazlan,  N.  S.  B.:  Numerical  simulation  of  the  effect  of  heavy  groundwater  abstraction  on 890 groundwater‐surface water interaction in Langat Basin, Selangor, Malaysia, Environmental Earth Sciences, 71, 891 1239‐1248, 10.1007/s12665‐013‐2527‐4, 2014. 892 

Page 29: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

29  

Molina‐Navarro,  E.,  Andersen,  H.  E.,  Nielsen,  A.,  Thodsen,  H.,  and  Trolle,  D.:  The  impact  of  the  objective 893 function in multi‐site and multi‐variable calibration of the SWAT model, Environmental Modelling & Software, 894 93, 255‐267, 10.1016/j.envsoft.2017.03.018, 2017. 895 Molina‐Navarro, E., Bailey, R. T., Andersen, H. E., Thodsen, H., Nielsen, A., Park, S., Jensen, J. S., Jensen, J. B., 896 and Trolle, D.: Comparison of abstraction scenarios simulated by SWAT and SWAT‐MODFLOW, Hydrological 897 Sciences Journal, 2019. 898 Moriasi, D. N., Gitau, M. W., Pai, N., and Daggupati, P.: Hydrologic and water quality models: Performance 899 measures and evaluation criteria, Transactions of the ASABE, 58, 1763‐1785, 2015. 900 NERI: Metadata for the area information system (in Danish: Metadata for Arealinformation Systemet). Danish 901 National Environmental Research Institute, Roskilde, 2000. 902 Nguyen, V. T., and Dietrich, J.: Modification of the SWAT model to simulate regional groundwater flow using 903 a multicell aquifer, Hydrological Processes, 32, 939‐953, 10.1002/hyp.11466, 2018. 904 Nielsen, A., Trolle, D., Me, W., Luo, L., Han, B.‐P., Liu, Z., Olesen, J. E., and Jeppesen, E.: Assessing ways to 905 combat eutrophication in a Chinese drinking water reservoir using SWAT, Marine and Freshwater Research, 906 64, 475, 10.1071/mf12106, 2013. 907 Niswonger, R. G., Panday, S., and Ibaraki, M.: MODFLOW‐NWT, a Newton formulation for MODFLOW‐2005, 908 US Geological Survey Techniques and Methods, 6, 44, 2011. 909 Pardo, I., and Garcia, L.: Water abstraction in small lowland streams: Unforeseen hypoxia and anoxia effects, 910 Science of the Total Environment, 568, 226‐235, 10.1016/j.scitotenv.2016.05.218, 2016. 911 Park,  S.:  Enhancement  of  Coupled  Surface/Subsurface  Flow  Models  in  Watersheds:  Analysis,  Model 912 Development, Optimization, and User Accessibility, Colorado State University, 2018. 913 Park, S., Nielsen, A., Bailey, R. T., Trolle, D., and Bieger, K.: A QGIS‐based graphical user interface for application 914 and  evaluation  of  SWAT‐MODFLOW  models,  Environmental  Modelling  &  Software, 915 https://doi.org/10.1016/j.envsoft.2018.10.017, 2018. 916 Parkin, G., Birkinshaw, S. J., Younger, P. L., Rao, Z., and Kirk, S.: A numerical modelling and neural network 917 approach to estimate the impact of groundwater abstractions on river flows, Journal of Hydrology, 339, 15‐918 28, 10.1016/j.jhydrol.2007.01.041, 2007. 919 Pfannerstill,  M.,  Guse,  B.,  and  Fohrer,  N.:  A  multi‐storage  groundwater  concept  for  the  SWAT  model  to 920 emphasize nonlinear groundwater dynamics in lowland catchments, Hydrological Processes, 28, 5573‐5612, 921 10.1002/hyp.10062, 2014. 922 Poulsen,  J.  B.:  Stream  flow‐its  estimation,  uncertainty  and  interaction with  groundwater  and  floodplains, 923 Aarhus University, 2013. 924 Sanz, D., Castano, S., Cassiraga, E., Sahuquillo, A., Gomez‐Alday, J. J., Pena, S., and Calera, A.: Modeling aquifer‐925 river interactions under the influence of groundwater abstraction in the Mancha Oriental System (SE Spain), 926 Hydrogeol J, 19, 475‐487, 10.1007/s10040‐010‐0694‐x, 2011. 927 Semiromi, M. T., and Koch, M.: Analysis of spatio‐temporal variability of surface–groundwater interactions in 928 the Gharehsoo river basin, Iran, using a coupled SWAT‐MODFLOW model, Environmental Earth Sciences, 78, 929 201, 2019. 930 Shafeeque, M., Cheema, M. J. M., Sarwar, A., and Hussain, M. W.: Quantification of Groundwater Abstraction 931 Using  Swat  Model  in  Hakra  Branch  Canal  System  of  Pakistan,  Pak  J  Agr  Sci,  53,  249‐255,  Doi 932 10.21162/Pakjas/16.4199, 2016. 933 Sith,  R.,  Watanabe,  A.,  Nakamura,  T.,  Yamamoto,  T.,  and  Nadaoka,  K.:  Assessment  of  water  quality  and 934 evaluation of best management practices in a small agricultural watershed adjacent to Coral Reef area in Japan, 935 Agricultural Water Management, 213, 659‐673, https://doi.org/10.1016/j.agwat.2018.11.014, 2019. 936 Stefania, G. A., Rotiroti, M., Fumagalli, L., Simonetto, F., Capodaglio, P., Zanotti, C., and Bonomi, T.: Modeling 937 groundwater/surface‐water  interactions  in  an  Alpine  valley  (the  Aosta  Plain,  NW  Italy):  the  effect  of 938 groundwater abstraction on surface‐water resources, Hydrogeol J, 26, 147‐162, 10.1007/s10040‐017‐1633‐x, 939 2018. 940 Surinaidu,  L.,  Gurunadha  Rao,  V.  V.  S.,  Srinivasa  Rao, N.,  and  Srinu,  S.:  Hydrogeological  and  groundwater 941 modeling studies  to estimate  the groundwater  inflows  into  the coal Mines at different mine development 942 

Page 30: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

30  

stages  using  MODFLOW,  Andhra  Pradesh,  India,  Water  Resources  and  Industry,  7‐8,  49‐65, 943 10.1016/j.wri.2014.10.002, 2014. 944 Thodsen, H., Andersen, H. E., Blicher‐Mathiesen, G., and Trolle, D.: The combined effects of fertilizer reduction 945 on high risk areas and increased fertilization on low risk areas, investigated using the SWAT model for a Danish 946 catchment,  Acta  Agriculturae  Scandinavica,  Section  B  —  Soil  &  Plant  Science,  65,  217‐227, 947 10.1080/09064710.2015.1010564, 2015. 948 Vainu,  M.,  and  Terasmaa,  J.:  The  consequences  of  increased  groundwater  abstraction  for  groundwater 949 dependent closed‐basin lakes in glacial terrain, Environmental Earth Sciences, 75, ARTN 9210.1007/s12665‐950 015‐4967‐5, 2016. 951 Wei, X., Bailey, R. T., Records, R. M., Wible, T. C., and Arabi, M.: Comprehensive simulation of nitrate transport 952 in  coupled  surface‐subsurface  hydrologic  systems  using  the  linked  SWAT‐MODFLOW‐RT3D  model, 953 Environmental Modelling & Software, 10.1016/j.envsoft.2018.06.012, 2018. 954 Yi, L., and Sophocleous, M.: Two‐way coupling of unsaturated‐saturated flow by  integrating the SWAT and 955 MODFLOW models with application in an irrigation district in arid region of West China, J Arid Land, 3, 164‐956 173, 2011. 957 Zhang,  X.,  Ren,  L.,  and  Kong,  X.:  Estimating  spatiotemporal  variability  and  sustainability  of  shallow 958 groundwater in a well‐irrigated plain of the Haihe River basin using SWAT model, Journal of Hydrology, 541, 959 1221‐1240, 10.1016/j.jhydrol.2016.08.030, 2016. 960 Zhulu, L.: Getting Started with PEST, Athens, The University of Georgia, 2010. 961 Zipper, S. C., Gleeson, T., Kerr, B., Howard, J. K., Rohde, M. M., Carah, J., and Zimmerman, J.: Rapid and accurate 962 estimates of streamflow depletion caused by groundwater pumping using analytical depletion functions, 2018. 963 

964  965  966  967  968  969  970  971  972  973  974  975  976  977  978  979  980  981  982  983  984  985  986  987  988  989  990  991  992  993 

Page 31: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

31  

Table 1. Farm types and crop rotations used to describe agricultural management in the Uggerby 994 

River Catchment (W: winter, S: spring). 995 

996 

997 

998 

999 

1000 

1001 

1002 

1003 

1004 

1005 

1006 

1007 

1008 

1009 

1010 

1011 

1012 

1013 

1014 

1015 

1016 

1017 

1018 

1019 

1020 

1021 

1022 

1023 

Rotation scheme Rotation type Farm

Type Manure N (kg N/ha)

% Farm area

Year 1 Year 2 Year 3 Year 4 Year 5

Agricultural land 1 Mixed and horticulture

<50 31.0 W. wheat W. wheat S. barley W. rape S. barley

Agricultural land 2 Dairy/Cattle 85-170

35.7

S. barley Grass S. barley Grass Grass

Agricultural land 3 Dairy/Cattle 85-170

33.3

S. barley S. barley W. wheat Corn silage Corn silage

Page 32: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

32  

Table 2. Initial ranges and calibrated values of the selected parameters for SWAT calibration. 1024 

1025 

Note: v_ means that the existing parameter value is to be replaced by a given value; r_ means that an existing parameter 1026 value is multiplied by (1+ a given value). 1027 

1028 

1029 

1030 

1031 

1032 

1033 

1034 

1035 

1036 

Parameter Description Initial range

Calibrated values

Subbasins:

4,5,7-13 (upstream)

Subbasins: 1,3,6,14-19

(downstream) v__SFTMP.bsn Snowfall temperature (°C) -1 – 1 0.175 v__SMFMN.bsn Melt factor for snow on December 21 (mm H2O °C-1

d-1) 1 – 2 1.287

v__SMFMX.bsn Melt factor for snow on June 21 (mm H2O °C-1 d-1) 1.6 – 3.5 2.467 v__SMTMP.bsn Snow melt base temperature (°C) -2.3 – 1 -1.342 v__SURLAG.bsn Surface runoff lag coefficient 1 – 10 6.379 v__ALPHA_BF.gw Baseflow alpha factor for shallow aquifer (l d-1) 0 – 1 0.453 0.639 v__ALPHA_BF_D.gw Baseflow alpha factor for deep aquifer (l d-1) 0 – 1 0.756 0.913 v__ALPHA_BNK.rte Baseflow alpha factor for bank storage (l d-1) 0 – 1 0.912 0.533 v__CH_K2.rte Effective hydraulic conductivity in main channel

alluvium (mm h-1) 0 – 75 57.068 45.018

r__CN2.mgt Initial SCS runoff curve number for moisture condition II

-0.3 – 0.3 -0.279 0.137

r__DDRAIN.mgt Depth to subsurface drain (mm) -0.3 – 0.3 0.066 -0.129 v__EPCO.hru Plant uptake compensation factor 0.01 – 1 0.163 0.254 v__ESCO.hru Soil evaporation compensation factor 0 – 1 0.466 0.931 r__GDRAIN.mgt Drain tile lag time (h) -0.3 – 0.3 0.052 -0.021 v__GWQMN.gw Threshold depth of water in the shallow aquifer

required for return flow to occur (mm) 0 – 2000 1435.04 960.32

v__GW_DELAY.gw Groundwater delay time (d) 0 – 200 116.28 123.40 v__GW_REVAP.gw Groundwater “revap” coefficient 0.02 – 0.1 0.092 0.0313 r__OV_N.hru Manning´s “n” value for overland flow -0.2 – 0.2 -0.037 -0.025 v__REVAPMN.gw Threshold depth of water in the shallow aquifer for

“revap” or percolation to the deep aquifer to occur (mm)

1000 – 2000

1633.81 1521.80

r__SOL_AWC().sol Available water capacity of the soil layer (mm H2O mm soil-1)

-0.8 – 0.8 -0.674 0.786

r__SOL_BD().sol Moist bulk density (g cm-3) -0.2 – 0.2 -0.067 0.156 r__SOL_K().sol Saturated hydraulic conductivity (mm h-1) -0.8 – 2 1.290 1.831 r__TDRAIN.mgt Time to drain soil to field capacity (h) -0.3 – 0.3 -0.097 -0.210 v__RCHRG_DP.gw Deep aquifer percolation fraction 0 – 0.4 0.296 0.219 AUTO_WSTRS Water stress threshold that triggers irrigation (mm) 70 30, 40, 60

Page 33: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

33  

Table 3. Initial values, ranges, and calibrated values of the selected parameters for SWAT-1037 

MODFLOW calibration using PEST. 1038 

Parameter Description Initial value

Parameter ranges Calibrated values

v__SFTMP.bsn --- 0.175 -0.946–0.351 0.351

v__SMFMN.bsn --- 1.287 1.117–1.424 1.424

v__SMFMX.bsn --- 2.467 2.387–3.129 2.387

v__SMTMP.bsn --- -1.342 -1.687– -0.46 -0.46

v__SURLAG.bsn --- 6.379 4.452– 8.151 4.964

v__ALPHA_BNK.rtea --- 0.912 0.7– 1 0.7

v__ALPHA_BNK.rteb --- 0.533 0.206– 0.617 0.231

v__CH_K2.rtea --- 57.068 29.322– 59.779 59.779

v__CH_K2.rteb --- 45.018 30.246– 60.088 41.182

r__CN2.mgta --- -0.279 -0.3– -0.106 -0.3

r__CN2.mgtb --- 0.137 -0.019– 0.175 0.0004

v__EPCO.hrua --- 0.163 0.077– 0.436 0.436

v__EPCO.hrub --- 0.255 0.01– 0.334 0.304

v__ESCO.hrua --- 0.466 0.227– 0.681 0.227

v__ESCO.hrub --- 0.931 0.684– 1 0.943

r__OV_N.hrua --- -0.037 -0.2– -0.02 -0.02

r__OV_N.hrub --- -0.025 -0.155– -0.005 -0.023

r__SOL_AWC().sola --- -0.675 -0.8– -0.316 -0.508

r__SOL_AWC().solb --- 0.786 0.344– 0.8 0.8

r__SOL_BD().sola --- -0.067 -0.187– -0.05 -0.185

r__SOL_BD().solb --- 0.156 0.077– 0.2 0.172

r__SOL_K().sola --- 1.29 0.902– 2 0.902

r__SOL_K().solb --- 1.831 1.012– 2 1.012

COND_1 Drain conductance 0.00467 0.00311 – 0.00622 0.00543

COND_2 Drain conductance 0.02487 0.01658 – 0.03316 0.03316 HK_CLAY Hydraulic conductivity of clay (m s-1) 3.84E-08 1E-09 – 4.4E-08 2.2E-08

HK_SILT Hydraulic conductivity of silt (m s-1) 5.00E-07 1E-07 – 9E-07 1E-07

HK_SS Hydraulic conductivity of silty sand (m s-1) 6.70E-06 1.51E-06 – 7.50E-06 7.5E-06

HK_SSCS Hydraulic conductivity of silty sand and clean sand (m s-1)

1.79E-05 1E-05 – 8E-05 1.79E-05

HK_CS Hydraulic conductivity of clean sand (m s-1) 0.000327 1E-04– 5E-04 3.15E-04

SS_CLAY Specific storage of clay (m-1) 0.001099 9.19E-04 – 1.28E-03 1.28E-03

SS_SILT Specific storage of silt (m-1) 0.000755 4.92E-04 – 1.02E-03 1.02E-03

SS_SAND Specific storage of sand (m-1) 0.000166 1.28E-04 – 2.03E-04 2.03E-04

SY_CLAY Specific yield of clay (%) 0.06 0.04 – 0.08 0.04

SY_SILT Specific yield of silt (%) 0.2 0.15 – 0.25 0.22

SY_SAND Specific yield of sand (%) 0.32 0.25 – 0.35 0.35

AUTO_WSTRS --- 30, 40, 60 30, 40, 60, 80

Notes: “a” means that the parameter applies to the upstream areas, including subbasins: 4, 5, 7-13, while “b” applies to 1039 

downstream areas, including subbasins 1, 3, 6, 14-19. “---” indicates that the corresponding parameters can be found in 1040 

Table 2. 1041 

1042 

1043 

Page 34: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

34  

Table 4. The summary statistics for the calibrated MODFLOW performance. 1044 

Layer number The number of observed heads

ME (Mean error)

MAE (Mean absolute error)

RMSE (Root mean squared error)

Layer 1 453 -0.59 1.94 2.84 Layer 3 572 -0.54 2.36 3.15 Layer 5 38 -1.24 3.44 5.00

All 1063 -0.59 2.22 3.11 1045 

1046 

1047 

1048 

1049 

1050 

1051 

1052 

1053 

1054 

1055 

1056 

1057 

1058 

1059 

1060 

1061 

1062 

1063 

1064 

1065 

1066 

1067 

1068 

1069 

1070 

1071 

1072 

1073 

1074 

1075 

Page 35: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

35  

Table 5. Performance statistics indices for daily runoff at the outlets of subbasin 13 and subbasin 18 1076 

during the calibration (2001-2008) and validation (2009-2015, in brackets) periods by SWAT, 1077 

SWAT-MODFLOW without PEST calibration, and SWAT-MODFLOW with PEST calibration. 1078 

1079 Outlets Used models PBIAS NSE R2

Subbasin 13 outlet

SWAT -3.9 (5.9) 0.66 (0.50) 0.67 (0.53) SWAT-MODFLOW

without PEST calibration

-6.9 (1.7)

0.72 (0.51)

0.75 (0.60) SWAT-MODFLOW with

PEST calibration

1.9 (9.4)

0.78 (0.54)

0.78 (0.61)

Subbasin 18 outlet

SWAT 2.0 (12.4) 0.74 (0.47) 0.74 (0.53) SWAT-MODFLOW

without PEST calibration

1.0 (11.0)

0.77 (0.46)

0.79 (0.57) SWAT-MODFLOW with

PEST calibration

3.3 (13.1)

0.81 (0.53)

0.82 (0.60) 1080 

1081 

1082 

1083 

1084 

1085 

1086 

1087 

1088 

1089 

1090 

1091 

1092 

1093 

1094 

1095 

1096 

1097 

1098 

1099 

1100 

1101 

1102 

1103 

1104 

1105 

1106 

1107 

1108 

1109 

1110 

1111 

1112 

1113 

1114 

1115 

Page 36: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

36  

Table 6. Summary statistics for the SWAT-MODFLOW calibration result. 1116  1117 

Observation group

Number of observed

data

Weight of observed data

Contribution to squared weighted residuals before

calibration by PEST

Contribution to squared weighted residuals after

calibration by PEST Streamflow A 2557 2 4410.3 3479.7 Streamflow B 2557 1 4911.7 4025.3

Well A 570 1 113 154.9 Well B 961 1 946.6 908.6

Sum 6645 --- 10381 8568.5 1118  1119  1120  1121  1122  1123 

1124  1125  1126  1127  1128  1129  1130  1131  1132  1133  1134  1135  1136  1137  1138  1139  1140  1141  1142 

1143 

1144 

1145 

1146 

1147 

1148 

1149 

1150 

1151 

1152 

1153 

1154 

1155 

1156 

1157 

1158 

Page 37: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

37  

Table 7. Average annual summary of the main components in the hydrological cycle of the Uggerby 1159 

River Catchment during the study period (2002-2015) simulated by SWAT and SWAT-MODFLOW, 1160 

respectively. 1161  1162 

Components SWAT SWAT-MODFLOW Precipitation (mm yr-1) 923 923 Surface flow (mm yr-1) 22 30

Lateral subsurface flow (mm yr-1) 89 64 Tile drain flow (mm yr-1) 20 0

Drain (MODFLOW, mm yr-1) 0 268 Groundwater flow (mm yr-1) 257 22 Total water yield (mm yr-1) 388 384

Actual evapotranspiration (mm yr-1) 503 516 Potential evapotranspiration (mm yr-1) 727 726

Soil storage (mm yr-1) 32 22 Average annual irrigation amount in the irrigated

HRUs (mm yr-1) 137 133

 1163  1164  1165  1166  1167  1168  1169  1170  1171  1172  1173  1174  1175  1176  1177  1178  1179  1180  1181  1182  1183  1184  1185  1186  1187  1188  1189  1190  1191  1192  1193 

1194 

1195 

1196 

Page 38: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

38  

Table 8. Average annual stream flow change (2002-2015) at subbasin 13 outlet and subbasin 18 1197 

outlet for each abstraction scenario from no-wells scenario and the corresponding annual abstraction 1198 

simulated in SWAT and SWAT-MODFLOW 1199  1200 

Scenarios

Scenario 2 (Only drinking water

wells)

Scenario 3 (Only irrigation wells)

Scenario 4 (Both drinking water and

irrigation wells) Model SWAT SWAT-

MODFLOW SWAT SWAT-

MODFLOW SWAT SWAT-

MODFLOW Average annual

stream flow decrease(-) or

increase(+) (106 m3 yr-1)

Subbasin 13 outlet

-0.024 -1.10 0.61 0.24 0.59 -0.73

Subbasin 18 outlet

-0.12 -2.53 1.60 -0.55 1.48

-1.79

Annual

abstraction (106 m3 yr-1)

Subbasins 4-5, 7-13

1.10

1.28

17.86

19.45

18.96

20.73

The entire cathment excluding

subbasin 19

4.01

3.96

40.54

39.26

44.55

43.22

 1201 Notes: Subbasin 13 outlet receives streamflow from subbasins 4-5, 7-13; Subbasin 18 outlet receives streamflow 1202 

from the entire catchment excluding subbasin 19. 1203 

 1204  1205  1206  1207  1208  1209  1210  1211  1212  1213  1214  1215  1216  1217  1218  1219  1220  1221  1222  1223  1224  1225  1226  1227  1228  1229 

Page 39: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

39  

Figure captions 1230 

Figure 1. Location of the Uggerby catchment and its delineation in SWAT, including subbasins 1231 

division, stream network definition based on the digital elevation model (DEM), hydrological 1232 

monitoring stations, and basin outlet. 1233 

Figure 2. The distribution and proportion of each land use (a), soil type (b), and slope band (c) after 1234 

reclassfication for HRU definition in SWAT. 1235 

Figure 3. SWAT and MODFLOW set-up coverage and the well locations distributed inside or outside 1236 

the Uggerby River Catchment. 1237 

Figure 4. Schematic representation of water transport routes in stream-aquifer system as simulated by 1238 

SWAT-MODFLOW, showing SWAT (green) and MODFLOW (blue) simulation processes. Adapted 1239 

from (Molina-Navarro et al., 2019). 1240 

Figure 5. Schematic diagram of the PEST optimization process. The “*” means file name or file path. 1241 

Figure 6. The simulated and observed head contours as well as the locations of observed wells within 1242 

layer 1 (a) and layer 3 (b), respectively. 1243 

Figure 7. Hydrographs of precipitation, observed and best simulated daily streamflow at the outlets of 1244 

subbasin 13 (station A) and subbasin 18 (station B) during the calibration period (2002-2008) and the 1245 

validation period (2009-2015) based on SWAT and SWAT-MODFLOW. The value in bracket is the 1246 

discharge on 16 October, 2014, which is outside the range of the plot area. 1247 

Figure 8. The sensitivity ranking of parameters in SWAT (a) and SWAT-MODFLOW (b) during 1248 

calibration. The composite sensitivity of parameters was calculated based on the Jacobian matrix and 1249 

the weight matrix after each PEST iteration and generated as an output once the PEST calibration was 1250 

finished. The composite sensitivity values vary a little among the different iterations. The average 1251 

value of each parameter among the 10 iterations for calibration is shown in the figure. More details 1252 

regarding composite parameter sensitivity can be found in (Doherty, 2018). The “a” indicates that the 1253 

parameter applies to the upstream areas, including subbasins: 4, 5, 7-13, and “b” indicates that he 1254 

parameter applies to downstream areas, including subbasins 1, 3, 6, 14-19. 1255 

Page 40: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

40  

Figure 9. The simulated groundwater heads for the first layer (a) and third layer (b) at initial conditions, 1256 

end of calibration period, and end of validation period by the calibrated SWAT-MODFLOW. 1257 

Figure 10. Hydrograph of daily simulated and observed groundwater heads (m a.s.l) of the two wells 1258 

located in layer 1 used for calibrating the variation of groundwater heads simulated by SWAT-1259 

MODFLOW where relatively continuous observed data is available. Also shown are summary 1260 

performance statistics. 1261 

Figure 11. Average annual water yield (total flow) (2002-2015) simulated for the scenarios (no wells, 1262 

scenario 1; only drinking water wells, scenario 2; only irrigation wells, scenario 3; both drinking water 1263 

and irrigation wells, scenario 4) with SWAT (a) and SWAT-MODFLOW (b) and divided into flow 1264 

components (Q = flow; GW = groundwater; AQ = aquifer). 1265 

Figure 12. The simulated daily streamflow in the no-wells scenario and daily discharge differences 1266 

between the abstraction scenarios (only drinking water wells, scenario 2; only irrigation wells, scenario 1267 

3; both drinking water and irrigation wells, scenario 4) and the no-wells scenario (scenario 1) at the 1268 

outlets of subbasin 13 (station A) and subbasin 18 (station B) during the entire study period (2002-1269 

2015) based on SWAT and SWAT-MODFLOW, respectively. The value 1.54 m3 s-1 in brackets is the 1270 

streamflow difference between the no-wells scenario and the scenario with only drinking water wells 1271 

on 24 March, 2010, which is outside the range of the plot area. 1272 

Figure 13. The hydrograph of simulated daily groundwater discharge to the stream network in the no-1273 

wells scenario and daily groundwater discharge differences between the abstraction scenarios (only 1274 

drinking water wells, scenario 2; only irrigation wells, scenario 3; both drinking water and irrigation 1275 

wells, scenario 4) and the no-wells scenario (scenario 1) in the upstream area of station A (a) and 1276 

upstream area of station B (b), respectively, during the entire study period (2002-2015), based on 1277 

SWAT-MODFLOW. 1278 

Page 41: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

41  

9

6

16

711

17

53

1

2

8

4

19

1314

15

1210

18

Sweden

Norway

Germany

Denmark

^̀ Hydrological stations and basin outlet

Streams

Subbasins

High : 108.4 m a.s.l

Low : -0.2 m a.s.l

0 150 30075 km

0 5 102.5 km

³A

B

 Figure 1. Location of the Uggerby catchment and its delineation in SWAT, including subbasins

division, stream network definition based on the digital elevation model (DEM), hydrological

monitoring stations, and basin outlet.

         

Page 42: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

42  

Land use

Agricultrulal land 1 (22.6 %)

Agricultrulal land 2 (19.6 %)

Agricultural land 3 (21.1 %)

Evergreen forest (11.3 %)

Pasture (11.2 %)

Residential area (7.6 %)

Wetland (6.6 %)

Soil types

DK2021 (7.3 %)

DK2028 (19 %)

DK2048 (7.6 %)

DK2211 (25 %)

DK2411 (7.6 %)

DK3011 (11.6 %)

DK998 (6.9 %)

DK999 (10 %)

Slope band

2% - 6% (29.1 %)

< 2% (48.7 %)

> 6% (22.2 %)

0 5 102.5 km

³(a) (b) (c)

Figure 2. The distribution and proportion of each land use (a), soil type (b), and slope band (c) after

reclassfication for HRU definition in SWAT.

                           

Page 43: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

43  

 Figure 3. SWAT and MODFLOW set-up coverage and the well locations distributed inside or outside

the Uggerby River Catchment.

                     

!.!.!.!.

!.

!.

!.!.!.

!.

!.

!. !.!.!.!.

!.!.

!.!.!.

!.!.

!.

!.

!.

!.

!.!.

!.!.

!.

!.

!.!.!.

!.!.

!.

!.!.!.

!.

!.

!.

!.!.!.!.!.

!.

!.

!.

!.!.

!.

!.

!.!.!.!.!.

!.

!.

!.

!.

!.

!.!.

!.!.

!.!.

!.!.

!.

!.

!.

!.!.

!.!.

!.

!.!.

!.

!. !.!.

!.

!.!.

!. !.!.

!.

!.

!.!.

!.

!.

!.

!.

!.

!.

!.

!.

!.!.

!.

!.!.

!.

!.

!.!.

!.

!.

!.

!.

!.

!.

!.

!.!.

!.

!.

!.!.

!.!.

!.

!.

!.

!.!.!. !.!.

!.

!.

!.

!.!.

!.

!.

!.

!.

!.

!.

!.

!.

!.

!.

!.

!.!.!.!.

!.

!. !.!.!.!.!.!.!.

!.

!.

!.!.!.

!.!.

!.

!.!.!. !.!.!.!.!.

!.!. !.!.

!.!.

!.!.

!.!.

!.

!.

!.!.!.!.!.!.

!.!. !.!.!.!. !.!.!. !.!.

!.!.!.!. !.!. !.!. !. !. !.!.!.!.!.!.!.!.!. !.!.

!.!.!.!.!.!.!.!.!.!. !.!.

!.!.!.!. !.!.!.!.!.

!. !.!. !.!.!.!.!.!.!.!.!.!.!. !.!. !.!. !.!. !. !.!.!. !.!.!.!. !.!. !. !.!.!.!. !. !.!.!.!.!.!.!.!. !.!.!. !.!.!. !.!. !.!.!. !.!. !.!. !.!.!. !. !.!.!. !. !.!. !.!.

!.!.!. !.!.!.!.!.!. !.!. !.!.

!.!.

!. !.!.!.

!.!.!.!. !.!.!.!.!. !.!.!.!.!. !.!. !.!. !.!.!.!. !.!. !. !.!.!. !. !.!.!.!.!.!. !.!.

!.!.!.!.!. !.!.

!.!. !.!.!.!.!.!. !.!.!.!. !. !.!.!. !. !.!.

!. !.!.

!.

!.!.

!.

9

6

16

711

17

53

1

2

8

4

19

1314

15

1210

18³

0 5 10 km

!. Wells outside

!. Irrigation wells

!. Drinking water wells

SWAT streams

SWAT subbasins

Active cells layer 1

Inactive cells layer 1

Page 44: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

44  

 Figure 4. Schematic representation of water transport routes in stream-aquifer system as simulated by

SWAT-MODFLOW, showing SWAT (green) and MODFLOW (blue) simulation processes. Adapted

from (Molina-Navarro et al., 2019).

Page 45: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

45  

                   

Figure 5. Schematic diagram of the PEST optimization process. The “*” means file name or file path.

PEST control file Parameter initial value and range, observed data

Input template file Surface water parameters (model.in.tpl) Groundwater parameters (*.drn. tpl, *.pval. tpl)

PEST executable file PEST optimization algorithm

Output instruction file

Model input model.in, *.drn, *.pval

Model batch file Swat_Edit.exe SWAT-MODFLOW.exe C: \Anaconda3 \python.exe *\ exsimvalue.py

Model output swatmf_out_MF_obs output.rch

Page 46: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

46  

 

Figure 6. The simulated and observed head contours as well as the locations of observed wells within

layer 1 (a) and layer 3 (b), respectively.

 

Page 47: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

47  

 

 

 

  

Figure 7. Hydrographs of precipitation, observed and best simulated daily streamflow at the outlets of

subbasin 13 (station A) and subbasin 18 (station B) during the calibration period (2002-2008) and the

validation period (2009-2015) based on SWAT and SWAT-MODFLOW. The value in bracket is the

discharge on 16 October, 2014, which is outside the range of the plot area.

0

20

40

60

80

100

120

1400

5

10

15

20

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Pre

cipi

tati

on (

mm

yr-1

)

Dis

char

ge (

m3

s-1)

(a) SWAT Station A - subbasin 13

ObservationSimulationPrecipitation

Calibration Validation( 17.1 m3 s-1 )

0

20

40

60

80

100

120

1400

5

10

15

20

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Pre

cipi

tati

on (

mm

yr-1

)

Dis

char

ge (

m3

s-1)

(b) SWAT-MODFLOW Station A - subbasin 13

( 22.1 m3 s-1 )

0

20

40

60

80

100

120

1400

5

10

15

20

25

30

35

40

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Pre

cipi

tati

on (

mm

yr-1

)

Dis

char

ge (

m3

s-1)

(c) SWAT Station B - subbasin 18

( 61.5 m3 s-1 )

0

20

40

60

80

100

120

1400

5

10

15

20

25

30

35

40

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Pre

cipi

tati

on (

mm

yr-1

)

Dis

char

ge (

m3

s-1)

(d) SWAT-MODFLOW Station B-subbasin 18

( 65.1 m3 s-1 )

Page 48: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

48  

 Figure 8. The sensitivity ranking of parameters in SWAT (a) and SWAT-MODFLOW (b) during

calibration. The composite sensitivity of parameters was calculated based on the Jacobian matrix and

the weight matrix after each PEST iteration and generated as an output once the PEST calibration was

finished. The composite sensitivity values vary a little among the different iterations. The average

value of each parameter among the 10 iterations for calibration is shown in the figure. More details

regarding composite parameter sensitivity can be found in (Doherty, 2018). The “a” indicates that the

parameter applies to the upstream areas, including subbasins: 4, 5, 7-13, and “b” indicates that the

parameter applies to downstream areas, including subbasins 1, 3, 6, 14-19.

0.000.050.100.150.200.250.300.350.400.450.500.550.600.650.700.750.800.850.900.951.00

SO

L_A

WC

(..).sol_aE

SC

O.hru_b

GW

_DE

LA

Y.gw

_aE

SC

O.hru_a

SO

L_A

WC

(..).sol_bS

OL

_BD

(..).sol_bS

OL

_K(..).sol_b

GW

_DE

LA

Y.gw

_bE

PC

O.hru_a

CN

2.mgt_b

CN

2.mgt_a

CH

_K2.rte_b

EP

CO

.hru_bA

LP

HA

_BN

K.rte_b

GW

_RE

VA

P.gw

_bS

OL

_BD

(..).sol_aG

WQ

MN

.gw_b

SO

L_K

(..).sol_aD

DR

AIN

.mgt_a

RE

VA

PM

N.gw

_bT

DR

AIN

.mgt_a

GW

_RE

VA

P.gw

_aS

FT

MP

.bsnA

LP

HA

_BN

K.rte_a

DD

RA

IN.m

gt_bS

MT

MP

.bsnA

LP

HA

_BF

_D.gw

_aC

H_K

2.rte_aS

MF

MN

.bsnO

V_N

.hru_aT

DR

AIN

.mgt_a

GD

RA

IN.m

gt_aS

MF

MX

.bsnA

LP

HA

_BF

.gw_b

GD

RA

IN.m

gt_bG

WQ

MN

.gw_a

RC

HR

G_D

P.gw

_bS

UR

LA

G.bsn

AL

PH

A_B

F_D

.gw_b

RC

HR

G_D

P.gw

_aR

EV

AP

MN

.gw_a

OV

_N.hru_a

AL

PH

A_B

F.gw

_a

P-v

alue

Parameters

(a) SWAT

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

HK

_CL

AY

HK

_SIL

T

HK

_SS

HK

_SS

CS

SS

_SA

ND

HK

_CS

SS

_SIL

T

SS

_CL

AY

CN

2_b

CO

ND

_1

OV

_N_a

CO

ND

_2

SO

L_B

D_b

SY

_CL

AY

OV

_N_b

SO

L_A

WC

_b

SO

L_B

_a

SY

_SA

ND

SY

_SIL

T

ES

CO

_a

CN

2_a

EP

CO

_a

SO

L_A

WC

_a

ES

CO

_b

AL

PH

A_b

EP

CO

_b

SF

TM

P

SM

TM

P

SO

L_K

_a

AL

PH

A_a

SO

L_K

_b

SM

FM

N

SM

FM

X

SU

RL

AG

CH

_K2_a

CH

_K2_b

The

com

posi

te p

aram

eter

sen

stiv

ity

Parameters

(b) SWAT-MODFLOW

Page 49: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

49  

Figure 9. The simulated groundwater heads for the first layer (a) and third layer (b) at initial conditions,

end of calibration period, and end of validation period by the calibrated SWAT-MODFLOW.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

¯

0 10 205 km

Initial Condition (01/01/1997) End of calibration period (31/12/2008) End of validation period (31/12/2015)

Layer_1 heads

88 m a.s.l

-1 m a.s.l

Layer_3 heads

67 m a.s.l

0 m a.s.l

(a)

(b)

Page 50: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

50  

 

 Figure 10. Hydrograph of daily simulated and observed groundwater heads (m a.s.l) of the two wells

located in layer 1 used for calibrating the variation of groundwater heads simulated by SWAT-

MODFLOW where relatively continuous observed data is available. Also shown are summary

performance statistics.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20

22

24

26

28

3/7/2007 3/7/2008 3/7/2009 3/7/2010 3/7/2011 3/7/2012 3/7/2013 3/7/2014 3/7/2015

Wat

er ta

ble

elev

atio

n (m

a.s

.l)(a) Well A

obs sims

R2 = 0.42PBIAS = 3.66

Calibration

20

22

24

26

8/19/2004 8/19/2005 8/19/2006 8/19/2007 8/19/2008 8/19/2009

Wat

er ta

ble

elev

atio

n (m

a.s

.l)

(b) Well B

R2 = 0.68PBIAS = -2.35

R2 = 0.84PBIAS = -1.91

Calibration Validation

R2 = 0.61 PBIAS = 0.80

Validation

Page 51: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

51  

 

 Figure 11. Average annual water yield (total flow) (2002-2015) simulated for the scenarios (no wells,

scenario 1; only drinking water wells, scenario 2; only irrigation wells, scenario 3; both drinking water

and irrigation wells, scenario 4) with SWAT (a) and SWAT-MODFLOW (b) and divided into flow

components (Q = flow; GW = groundwater; AQ = aquifer).

 

 

 

 

21.8 21.8 21.9 21.9

87.7 87.7 89.3 89.3

19.6 19.6 19.6 19.6

177.8 177.5 179.6 179.3

76.6 76.6 77.8 77.8

0.0

50.0

100.0

150.0

200.0

250.0

300.0

350.0

400.0

450.0

No wells Only_drink Only_irri Both

Flo

w (

mm

yr-1

)

(a) SWAT

Surface Q Lateral Q Tile Q GW (shallow AQ) Q GW (deep AQ) Q

29.9 29.9 29.9 29.9

62.9 62.9 64.0 63.7

273.8 268.3 274.1 267.8

21.9 21.7 21.8 21.6

0.0

50.0

100.0

150.0

200.0

250.0

300.0

350.0

400.0

450.0

No wells Only_drink Only_irri Both

Flo

w (

mm

yr-1

)

(b) SWAT-MODFLOW

Surface Q Lateral Q Drain Q GW Q

Page 52: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

52  

 

 

 

 Figure 12. The simulated daily streamflow in the no-wells scenario and daily discharge differences

between the abstraction scenarios (only drinking water wells, scenario 2; only irrigation wells, scenario

3; both drinking water and irrigation wells, scenario 4) and the no-wells scenario (scenario 1) at the

outlets of subbasin 13 (station A) and subbasin 18 (station B) during the entire study period (2002-

2015) based on SWAT and SWAT-MODFLOW, respectively. The value 1.54 m3 s-1 in brackets is the

streamflow difference between the no-wells scenario and the scenario with only drinking water wells

on 24 March, 2010, which is outside the range of the plot area.

0

2

4

6

8

10

12

14

16

18

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Dis

char

ge (

m3

s-1)

Dis

char

ge d

iffe

renc

e (m

3s-1

)

(a) SWAT Station A - subbasin 13 outlet

Only drinking wells Only irrigation wellsBoth drinking and irrigation wells No_wells_discharge

0

5

10

15

20

25

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Dis

char

ge (

m3

s-1)

Dis

char

ge d

iffe

renc

e (

m3

s-1)

(b) SWAT-MODFLOW Station A - subbasin 13 outlet

(1.54 m3 s-1)

0

10

20

30

40

50

60

70

-1

-0.5

0

0.5

1

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Dis

char

ge (

m3

s-1)

Dis

chag

e di

ffer

ence

(m

3 s-1

)

(c) SWAT Station B- subbasin 18 outlet

0

10

20

30

40

50

60

70

-1

-0.5

0

0.5

1

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Dis

char

ge (

m3

s-1)

Dis

hcha

rge

diff

eren

ce (

m3

s-1)

(d) SWAT-MODFLOW Station B - subbasin 18 outlet

‐6‐14

10/15/2014 10/18/2014 10/21/2014

Extrem high flow period

Page 53: the impacts of groundwater abstractions for irrigation and … · 2019-08-31 · 44 abstraction scenarios analysis, both models indicated that abstraction for drinking water caused

53  

  

 Figure 13. The hydrograph of simulated daily groundwater discharge to the stream network in the no-

wells scenario and daily groundwater discharge differences between the abstraction scenarios (only

drinking water wells, scenario 2; only irrigation wells, scenario 3; both drinking water and irrigation

wells, scenario 4) and the no-wells scenario (scenario 1) in the upstream area of station A (a) and

upstream area of station B (b), respectively, during the entire study period (2002-2015), based on

SWAT-MODFLOW. 

 

0

5

10

15

20

25

30

35

40

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Groundwater discharge (m

3s‐1)

Gro

undw

ater

dis

char

ge d

iffe

renc

e (m

3s-1

)(a) SWAT - MODFLOW Station A - subbasin 13 outlet

only drinking wells only irrigation wellsBoth drinking and irrigation wells Groundwater discharge

0

5

10

15

20

25

30

35

40

45

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

1/1/2002 1/1/2004 1/1/2006 1/1/2008 1/1/2010 1/1/2012 1/1/2014 1/1/2016

Gro

undw

ater

dis

char

ge (

m3

s-1)

Gro

undw

ater

dis

char

ge d

iffe

renc

e (m

3s-1

)

(b) SWAT - MODFLOW Station B - subbasin 18 outlet